The DOI matches its Crossref record; 71 of the 88 references it cites were found in a registry or an index; 8 of the 27 sources its claims cite were checked for support.
Vision Transformers for Dense Prediction
Open accessAbstract
We introduce dense prediction transformers, an architecture that leverages vision transformers in place of convolutional networks as a backbone for dense prediction tasks. We assemble tokens from various stages of the vision transformer into image-like representations at various resolutions and progressively combine them into full-resolution predictions using a convolutional decoder. The transformer backbone processes representations at a constant and relatively high resolution and has a global receptive field at every stage. These properties allow the dense prediction transformer to provide finer-grained and more globally coherent predictions when compared to fully-convolutional networks. Our experiments show that this architecture yields substantial improvements on dense prediction tasks, especially when a large amount of training data is available. For monocular depth estimation, we observe an improvement of up to 28% in relative performance when compared to a state-of-the-art fully-convolutional network. When applied to semantic segmentation, dense prediction transformers set a new state of the art on ADE20K with 49.02% mIoU. We further show that the architecture can be fine-tuned on smaller datasets such as NYUv2, KITTI, and Pascal Context where it also sets the new state of the art. Our models are available at https://github.com/intel-isl/DPT.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
32 Own result · 23 Own interpretation · 11 No basis given · 10 No full text of the sources · 3 Supported · 2 Partly supported · 1 Supported so far · 1 Own method · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“Specifically, we use the recently proposed vision transformer (ViT) [11] as a backbone architecture.”
§ Vision Transformers for Dense Prediction > 1. Introduction
What the sources say
- Supportedcites [11] 11. Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. source
Section 3.1 (p. 3) introduces and describes the Vision Transformer (ViT) architecture. The citation identifies the ViT used as the citing authors’ backbone; their choice to use it is not a claim attributed to this source. (LLM, conf=0.93)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Virtually all existing architectures for dense prediction are based on convolutional networks [6,31,34,42,49,50,53].”
§ Vision Transformers for Dense Prediction > 1. Introduction
What the sources say
- Partly supportedcites [6] 6. Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L. Yuille. DeepLab: Semantic im- age segmentation with deep convolutional nets, atrous con- volution, and fully connected crfs. TPAMI, 40(4):834-848, 2018. sourceExperimental — accuracy not yet measured
The Introduction (p. 1) says convolutional models such as PixelCNN had received much more attention for pixel-by-pixel image generation. It does not establish that virtually all architectures across the broader category of dense prediction are convolutional; it also discusses recurrent models and proposes a self-attention alternative (Introduction, p. 2). (LLM, conf=0.78)
- Not checked — no full text of the sourcecites [31] 31. Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U- Net: Convolutional networks for biomedical image segmen- tation. In MICCAI, 2015. source
- Supportedcites [34] 34. Evan Shelhamer, Jonathan Long, and Trevor Darrell. Fully convolutional networks for semantic segmentation. CVPR, 2015. source
The paper is an example of a convolutional-network architecture for dense prediction: its Abstract (p. 1) describes predicting a depth map, and §3.1.2 (p. 3) identifies its convolutional layers and target-depth output. As one of several citations, it supports the sentence by exemplifying such an architecture, rather than establishing the prevalence claim on its own. (LLM, conf=0.88)
- Could not verifycites [42] 42. Jingdong Wang, Ke Sun, Tianheng Cheng, Borui Jiang, Chaorui Deng, Yang Zhao, Dong Liu, Yadong Mu, Mingkui Tan, Xinggang Wang, Wenyu Liu, and Bin Xiao. Deep high-resolution representation learning for visual recogni- tion. TPAMI, 2020. sourceExperimental — accuracy not yet measured
Section 2.1 (p. 8) describes transformer language-model architectures, not architectures for dense prediction. Searches for dense prediction, convolutional networks, and pixel-wise segmentation found no passage addressing whether dense-prediction architectures are predominantly convolutional. (LLM, conf=0.78)
- Not matchedcites [49] 49. Fisher Yu and Vladlen Koltun. Multi-scale context aggrega- tion by dilated convolutions. In ICLR, 2016.
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not checked — no full text of the sourcecites [50] 50. Yuhui Yuan, Xilin Chen, and Jingdong Wang. Object- contextual representations for semantic segmentation. In ECCV, 2020. source
- Not checked — no full text of the sourcecites [53] 53. Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing network. In CVPR, 2017. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“We notice that the biggest gains in performance for zero-shot transfer were achieved for datasets that feature dense, high-resolution evaluations [15,55,59].”
§ Vision Transformers for Dense Prediction > Additional information
What the sources say
- Not checked — no full text of the sourcecites [15] 15. Andreas Geiger, Philip Lenz, and Raquel Urtasun. Are we ready for autonomous driving? The KITTI vision benchmark suite. In CVPR, 2012. source
- Supportedcites [55] 55. D. J. Butler, J. Wulff, G. B. Stanley, and M. J. Black. A naturalistic open source movie for optical flow evaluation. In ECCV, 2012. source
The citation identifies MPI-Sintel as one of the datasets in the citing authors’ comparison; the performance gains are the citing authors’ observation, not a result attributed to this paper. Design Decisions and Comparison to Middlebury (p. 5) describes Sintel’s higher-resolution frames, and The Sintel Data Set (p. 9) describes its flow fields and pixel-level error evaluation. (LLM, conf=0.86)
- Not checked — no full text of the sourcecites [59] 59. Thomas Schöps, Johannes L. Schönberger, Silvano Galliani, Torsten Sattler, Konrad Schindler, Marc Pollefeys, and An- dreas Geiger. A multi-view stereo benchmark with high- resolution images and multi-camera videos. In CVPR, 2017. source
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Note that ResNext101-WSL was pretrained on a billion-scale corpus of weakly supervised data [25] in addition to ImageNet pretraining.”
§ Vision Transformers for Dense Prediction > Skip connections.
What the sources say
- Partly supportedcites [25] 25. Dhruv Mahajan, Ross Girshick, Vignesh Ramanathan, Kaiming He, Manohar Paluri, Yixuan Li, Ashwin Bharambe, and Laurens van der Maaten. Exploring the limits of weakly supervised pretraining. In ECCV, 2018. sourceExperimental — accuracy not yet measured
Section 2.3 (p. 4) identifies the models as ResNeXt-101, and Section A.5 (p. 23) says they were pretrained on billions of Instagram images using hashtags as labels. But it describes subsequent training on ImageNet as fine-tuning, not additional ImageNet pretraining; Section 3.2 (p. 12) also distinguishes Instagram-pretrained models from ImageNet-pretrained baselines. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“While ignoring the token yields good performance, projection provides slightly better performance on average.”
§ Vision Transformers for Dense Prediction > Skip connections.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Adding the token, on the other hand, yields worse performance than simply ignoring it.”
§ Vision Transformers for Dense Prediction > Skip connections.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Both DPT variants significantly outperform the state of the art.”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Note that both architectures have similar latency to MiDaS (Table 9).”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“ViT-Large outperforms all other backbones but is also almost three times larger than ViT-Base and ViT-Hybrid.”
§ Vision Transformers for Dense Prediction > Skip connections.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We again see that DPT can provide strong performance even on smaller datasets.”
§ Vision Transformers for Dense Prediction > ADE20K.
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Probably not a substantial claim. 11
- It may state two separate claims as one. 13
- The quote only partly states this claim. 1
Every source the paper cites for this claim was checked and supports it.
In the paper
“In particular, it has recently been demonstrated that a direct application of token-based transformer architectures that have been successful in NLP can yield competitive performance on image classification [11].”
§ Vision Transformers for Dense Prediction > 2. Related Work
What the sources say
- Supportedcites [11] 11. Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. source
The Introduction (p. 1) says image patches are treated like NLP tokens and fed to a standard Transformer. The Conclusion (p. 9) reports that this approach matches or exceeds the state of the art on many image-classification datasets when coupled with pre-training on large datasets. (LLM, conf=0.93)
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“A key insight of this work was that, like transformer models in NLP, vision transformers need to be paired with a sufficient amount of training data to realize their potential.”
§ Vision Transformers for Dense Prediction > 2. Related Work
What the sources say
- Supportedcites [11] 11. Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. source
The Introduction (p. 1) says ViT performs modestly on mid-sized datasets and does not generalize well with insufficient data, but attains excellent results when pre-trained at sufficient scale. Section 4.3 (p. 6) reports that larger ViT models show their full benefit only with the largest pre-training dataset tested. The Conclusion (p. 9) also describes the approach as working well when coupled with pre-training on large datasets. (LLM, conf=0.94)
plainconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Similarly, the ViT-Hybrid architecture extracts features at 1 16 the input resolution, which is twice as high as the lowest-resolution features that are commonly used with convolutional backbones.”
§ Vision Transformers for Dense Prediction > 3. Architecture
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“It contains about 1.4 million images and is, to the best of our knowledge, the largest training set for monocular depth estimation that has ever been compiled.”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“When applied to semantic segmentation, dense vision transformers set a new state of the art on ADE20K with 49.02% mIoU.”
§ Vision Transformers for Dense Prediction > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The quote only partly states this claim.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We observe that the DPT tends to produce cleaner and finer-grained delineations of object boundaries and that the predictions are also in some cases less cluttered.”
§ Vision Transformers for Dense Prediction > ADE20K.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“DPT can better reconstruct fine details while also improving global coherence in areas that are challenging for the convolutional architecture (for example, large homogeneous regions or relative depth arrangement across the image).”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The average relative improvement over the best published architecture, MiDaS, is more than 23% for DPT-Hybrid and 28% for DPT-Large.”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“DPT-Hybrid achieves this with a comparable network capacity (Table 9), while DPT-Large is about 3 times larger than MiDaS.”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We observe the tendency that attention is spatially more localized close to the reference token in shallow layers (leftmost columns), whereas deeper layers (rightmost columns) frequently attend across the whole image.”
§ Vision Transformers for Dense Prediction > Additional information
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“ViT-Base has comparable performance to ResNext101-WSL, while ViT-Hybrid and ViT-Large improve performance even though they have been pretrained on significantly less data.”
§ Vision Transformers for Dense Prediction > Skip connections.
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
In the paper
“While we observe a general trend of an improvement in per-class IoU in comparison to the baseline [51], we do not observe a strong pattern across classes.”
§ Vision Transformers for Dense Prediction > Additional information
What the sources say
- Not checked — no full text of the sourcecites [51] 51. Hang Zhang, Chongruo Wu, Zhongyue Zhang, Yi Zhu, Zhi Zhang, Haibin Lin, Yue Sun, Tong He, Jonas Muller, R. Manmatha, Mu Li, and Alexander Smola. ResNeSt: Split- attention networks. arXiv preprint arXiv:2004.08955, 2020.
hedgedcorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“While we observe a general trend of an improvement in per-class IoU in comparison to the baseline [51], we do not observe a strong pattern across classes.”
§ Vision Transformers for Dense Prediction > Additional information
What the sources say
- Not checked — no full text of the sourcecites [51] 51. Hang Zhang, Chongruo Wu, Zhongyue Zhang, Yi Zhu, Zhi Zhang, Haibin Lin, Yue Sun, Tong He, Jonas Muller, R. Manmatha, Mu Li, and Alexander Smola. ResNeSt: Split- attention networks. arXiv preprint arXiv:2004.08955, 2020.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We observe more details and also better global depth arrangement in DPT predictions when compared to the fullyconvolutional baseline.”
§ Vision Transformers for Dense Prediction > Additional information
plaincorrelational- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The transformer backbone processes representations at a constant and relatively high resolution and has a global receptive field at every stage.”
§ Vision Transformers for Dense Prediction > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As long as the image size is divisible by p, the embedding procedure can be applied and will produce a varying number of image tokens N p .”
§ Vision Transformers for Dense Prediction > 3. Architecture
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Unlike fully-convolutional networks, the vision transformer backbone foregoes explicit downsampling operations after an initial image embedding has been computed and maintains a representation with constant dimensionality throughout all processing stages.”
§ Vision Transformers for Dense Prediction > 1. Introduction
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Since transformers are known to realize their full potential only when an abundance of training data is available, monocular depth estimation is an ideal task to test the capabilities of DPT.”
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
boosteddirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This is in stark contrast to convolutional networks, which progressively increase their receptive field as features pass through consecutive convolution and downsampling layers.”
§ Vision Transformers for Dense Prediction > 3. Architecture
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“DPT-Large performs slightly worse, likely because of the significantly smaller dataset compared to our previous experiments.”
§ Vision Transformers for Dense Prediction > ADE20K.
hedgedconditional- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“DeIT trains the ViT architecture with a more data-efficient pretraining procedure.”
§ Vision Transformers for Dense Prediction > Skip connections.
What the sources say
- Not checked — no full text of the sourcecites [38] 38. Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé Jégou. Training data-efficient image transformers & distillation through at- tention. arXiv preprint arXiv:2012.12877, 2020. source
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Note that the DeIT-Base architecture is identical to ViT-Base, while DeIT-Base-Dist introduces an additional distillation token, which we ignore in the Reassemble operation.”
§ Vision Transformers for Dense Prediction > Skip connections.
What the sources say
- Not checked — no full text of the sourcecites [38] 38. Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, and Hervé Jégou. Training data-efficient image transformers & distillation through at- tention. arXiv preprint arXiv:2012.12877, 2020. source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Similar to prior work on transformers, DPT unfolds its full potential when trained on large-scale datasets.”
§ Vision Transformers for Dense Prediction > 5. Conclusion
plaindirect- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: fails — The paper may give no reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
What the paper did — it cites no source for this, so there is none to check. The quote shows it is in the paper.
In the paper
“Note that results for DPT and Mi-DaS are computed at the same input resolution (384 pixels).”
§ Vision Transformers for Dense Prediction > Additional information
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: fails — The paper’s conclusions may not depend on it.
- Based: passes — The paper gives a reason why it holds.
Probably not a substantial claim. The paper’s conclusions may not depend on it.
The paper relies on this without support.
The choice of backbone architecture has a large influence on the capabilities of the overall model, because information lost in the encoder cannot be recovered in the decoder.
“However, it is widely recognized that the choice of backbone architecture has a large influence on the capabilities of the overall model, as any information that is lost in the encoder is impossible to recover in the decoder.”
The paper relies on this without support.
Feature resolution and granularity are critical for dense prediction, where the architecture should ideally resolve features at or close to the input resolution.
“While feature resolution and granularity may not matter for some tasks, such as image classification, they are critical for dense prediction, where the architecture should ideally be able to resolve features at or close to the resolution of the input image.”
The paper relies on this without support.
Convolutions are linear operators with a limited receptive field.
“Convolutions, by definition, are linear operators that have a limited receptive field.”
The paper relies on this without support.
Downsampling intermediate representations is necessary to keep memory consumption at levels feasible with existing computer architectures.
“Downsampling the intermediate representations is necessary to keep memory consumption at levels that are feasible with existing computer architectures.”
The paper relies on this without support.
The transformer backbone maintains a constant feature resolution.
“Since the transformer backbone maintains a constant feature resolution, it is not clear at which points in the backbone features should be tapped.”
The paper relies on this without support.
Performance of fully-convolutional architectures suffers heavily when performing inference at an input resolution significantly different from the training resolution.
“Performance thus suffers heavily when performing inference at an input resolution that is significantly different from the training resolution.”
The paper relies on this without support.
Transformer encoders have a global receptive field in every layer.
“Transformer encoders, on the other hand, have a global receptive field in every layer.”
The paper relies on this without support.
The reduced meta-dataset datasets provide high-quality ground truth.
“We choose these datasets since they provide high-quality ground truth.”
The paper relies on this without support.
Convolutional architectures offer natural points of interest for passing features from the encoder to the decoder, namely before or after downsampling of the representation.
“Convolutional architectures offer natural points of interest for passing features from the encoder to the decoder, namely before or after downsampling of the representation.”
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Vision transformers, which are a deep learning approach to encoding low-dimensional input into a high dimensional feature space, have established new frontiers in depth estimation compared to convolutional neural networks (Ranftl et al., 2021).”
From Very high resolution canopy height maps from RGB imagery using self-supervised vision transformer and convolutional decoder trained on aerial lidar (2024) · cites this paper as Ranftl et al., 2021
In the citing paper
“However, the Transformer framework does not involve dimensionality reduction10,12, thus keeping all attention layer traceable to the original input features13, thus making the models interpretable.”
From Transformer for one stop interpretable cell type annotation (2023) · cites this paper as [10]
In the citing paper
“For example, the change of dimensions and non-linear aggregation of features throughout the autoencoder’s deep processing stages leads to untraceable and uninterpretable latent space and loss of information and feature resolution9,10.”
From Transformer for one stop interpretable cell type annotation (2023) · cites this paper as [10]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Vision Transformers for Dense Prediction > 4.1. Monocular Depth Estimation
“Our architecture matches or improves state-of-the-art performance on both datasets in all metrics.”
§ Vision Transformers for Dense Prediction > 5. Conclusion
“| BTS [20] | 0.956 | 0.993 | 0.998 | 0.059 2.756 | 0.096 | | DPT-Hybrid 0.959 | 0.995 | 0.999 | 0.062 2.573 | 0.092 |”
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 4×
“We perform a similar experiment with the hybrid architecture in Table 6 (bottom), where R0 and R1 refer to us-ing features from the first and second downsampling stages of the ResNet50 embedding network.”
MethodInfluentialVaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
U-Net: Convolutional Networks for Biomedical Image Segmentation
Ronneberger, O., Fischer, P., & Brox, T. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-24574-4_28 · Cited by 99,404 · cited 3×
“These include training at higher input resolution (if the computational budget permits), dilated convolutions [51] to rapidly increase the receptive field without downsampling, appropriately-placed skip connections from multiple stages of the encoder to the decoder [33], or, more recently, by connecting multiresolutio…”
BackgroundAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
ImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 5×
“The encoder is frequently based on an image classification network, also called the backbone, that is pretrained on a large corpus such as ImageNet [9].”
InfluentialBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 6×
“Following work in NLP, the ViT additionally adds a special token that is not grounded in the input image and serves as the final, global image representation which is used for classification.”
BackgroundMethodInfluentialAn Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472 · cited 8×
“On a high level, the vision transformer (ViT) [11] operates on a bag-of-words representation of the image [38].”
BackgroundMethodResultInfluentialFully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834 · cited 2×
“Virtually all existing architectures for dense prediction are based on convolutional networks [6, 33, 36, 44, 51, 52, 55].”
BackgroundLanguage Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Research works printed but not linked: 3
- yuan, 2020, Objectcontextual representations for semantic segmentation, ECCV
- godard, 2017, Unsupervised monocular depth estimation with leftright consistency, CVPR
- casser, 2019, Unsupervised learning of depth and ego-motion: A structured approach, AAAI
Cited by
Sorted by citations
Items 401–500 of 2,248
RADIOv2.5: Improved Baselines for Agglomerative Vision Foundation Models
Heinrich, G., Ranzinger, M., Yin, H.-X., Lu, Y., Kautz, J., Tao, A., Catanzaro, B., & Molchanov, P. (2024) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr52734.2025.02094 · Cited by 118
TrackFlow: Multi-Object Tracking with Normalizing Flows
Mancusi, G., Panariello, A., Porrello, A., Fabbri, M., Calderara, S., & Cucchiara, R. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00874 · Cited by 21
“The estimation of distances from monocular images is a crucial task for many computer vision applications and has been studied for many years [1, 29, 63, 30, 64, 47].”
BackgroundReview of monocular depth estimation methods
Zhang, Z., Zhang, Y., Li, Y., & Wu, L. (2025) · Journal of Electronic Imaging · doi:10.1117/1.jei.34.2.020901 · Cited by 23
“Ranftl et al. 122 used a CNN to extract feature maps at ð 1 ∕ 16 Þ th resolution, passed to a vision transformer for global information aggregation.”
Ghosh, S., Sarkar, S., Ghosh, S., et al. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05380-7 · Cited by 19
“The Vision Transformer (ViT) is a recent and noteworthy advancement in computer vision [18], gaining attention for its outstanding performance in various domains, especially classification tasks [19].”
Towards Robust Monocular Depth Estimation: A New Baseline and Benchmark
Xian, K., Cao, Z., Shen, C., & Lin, G. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-023-01979-4 · Cited by 20
“To further improve the generalizability of depth estimation models, recent methods (Lasinger et al., 2020; Yin et al., 2022, 2021; Ranftl et al., 2021) propose to train models on multiple in-the-wild RGB-D datasets simultaneously.”
InfluentialLightDepth: Single-View Depth Self-Supervision from Illumination Decline
Rodríguez-Puigvert, J., Batlle, V. M., Montiel, J. M. M., Martínez-Cantín, R., Fua, P., Tardós, J. D., & Civera, J. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01945 · Cited by 19
“To name a few, network architectures evolved to fully convolutional in Laina et al. [31] and more recently to transformers [48, 4, 34].”
InfluentialMapNeRF: Incorporating Map Priors into Neural Radiance Fields for Driving View Simulation
Wu, C., Sun, J., Shen, Z., et al. (2023) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros55552.2023.10341681 · Cited by 18
“Optionally, we can use DPT [47] to filter occluded regions on D out using least square fitting similar to [48].”
MethodSparsegs: Sparse View Synthesis Using 3D Gaussian Splatting
Xiong, H., Muttukuru, S., Xiao, H., Upadhyay, R., Chari, P., Zhao, Y., & Kadambi, A. (2025) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv66043.2025.00100 · Cited by 55
“We opt for the Marigold monocular depth estimator [10], though any pre-trained depth estimation model [17, 26, 27, 33] could work.”
Persistent Nature: A Generative Model of Unbounded 3D Worlds
Chai, L., Tucker, R. P., Li, Z., Isola, P., & Snavely, N. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01999 · Cited by 27
“We also obtain auxiliary outputs – disparity and sky segmentation – using the pretrained DPT [61] model.”
InfluentialSDPT: Semantic-Aware Dimension-Pooling Transformer for Image Segmentation
Cao, H., Chen, G., Zhao, H., Jiang, D., Zhang, X., Tian, Q. C., & Knoll, A. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3417813 · Cited by 19
“Previously, researchers have used pre-trained ViTs as encoders to improve segmentation performance [38], [44], [63].”
MethodENRICH: Multi-purposE dataset for beNchmaRking In Computer vision and pHotogrammetry
Marelli, D., Morelli, L., Farella, E. M., Bianco, S., Ciocca, G., & Fabio Remondino, A. (2023) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2023.03.002 · Cited by 21
3D Moments from Near-Duplicate Photos
Wang, Q., Li, Z., Salesin, D., Snavely, N., Curless, B., & Kontkanen, J. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00388 · Cited by 18
“Given an input image and its estimated monocular depth [31], we first apply ag-”
BackgroundMethodInfluentialLiu, Y., Yu, Y., Ouyang, J., Jiang, B., Ostmeier, S., Wang, J., Lu-Liang, S. Y., Yang, Y., Yang, G., Michel, P., et al. (2024) · Radiology · doi:10.1148/radiol.240137 · Cited by 28
Sun, L., Liu, H., Wang, X., et al. (2025) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2025.3572486 · Cited by 18
“Previous studies [47], [48] have highlighted the crucial role of large-receptive fields.”
ImplantFormer: vision transformer-based implant position regression using dental CBCT data
Yang, X., Li, X., Li, X., Wu, P., Shen, L., & Deng, Y. (2024) · Neural Computing and Applications · doi:10.1007/s00521-023-09411-1 · Cited by 23
“3, the reassemble block [43] firstly transform the multi-levels ViT features to multi-resolutions feature maps and then the image-like feature maps are integrated by upsampling and concatenation.”
MethodWeatherStream: Light Transport Automation of Single Image Deweathering
Zhang, H., Ba, Y., Yang, E., Mehra, V., Gella, B., Suzuki, A., Pfahnl, A. E., Chandrappa, C. C., Wong, A., & Kadambi, A. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01297 · Cited by 24
“Deweathering models trained on WeatherStream may also support the re-use of pretrained models for downstream tasks such as detection [39, 50], segmentation [5, 36, 52], depth prediction [1, 49, 65], and completion [23, 66].”
MethodLee, H., Savva, M., & Chang, A. X. (2024) · Computer Graphics Forum · doi:10.1111/cgf.15061 · Cited by 32
“To achieve this, SceneScape employs two pretrained models: Stable Diffusion (SD) inpainting and a depth estimation model [RBK21].”
MethodA review on 3D Gaussian splatting for sparse view reconstruction
Liu, H., Liu, B., Hu, Q., Du, P., Li, J., Bao, Y., & Wang, F. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11171-4 · Cited by 20
CAR: Class-Aware Regularizations for Semantic Segmentation
Huang, Y., Kang, D., Chen, L., Zhe, X., Jia, W., Bao, L., & He, X. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19815-1_30 · Cited by 19
Self-Supervised Monocular Depth Underwater
Amitai, S., Klein, I., & Treibitz, T. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161161 · Cited by 20
“a newer method [3], [28] use transformers to improve performance.”
MethodDual-Field-of-View Context Aggregation and Boundary Perception for Airport Runway Extraction
Jing, W. P., Yuan, Y., & Wang, Q. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3271676 · Cited by 19
“The experiments compare the performance differences between DCBP and SETR [32], DPT [47], Twins [48], Segmenter [33], Pool-Former [49], Segformer [34], and Swin [50].”
MethodBridging the Domain Gap between Synthetic and Real-World Data for Autonomous Driving
Bai, X., Luo, Y., Jiang, L., Gupta, A., Kaveti, P., Singh, H., & Ostadabbas, S. (2023) · ACM Journal on Autonomous Transportation Systems · doi:10.1145/3633463 · Cited by 19
“Additionally, existing unpaired style translation algorithms eliminate the need for manual labeling and ground truth annotation.”
MethodResultZero-shot visual reasoning through probabilistic analogical mapping
Webb, T. W., Fu, S., Bihl, T., Holyoak, K. J., & Lu, H. (2023) · Nature Communications · doi:10.1038/s41467-023-40804-x · Cited by 19
“Thisresultsuggeststhatthein fl uenceofobject-levelsimilarityin humananalogicalreasoningmay bedriveninpartbythefactthatitisa useful constraint in the context of complex, real-world analogy-making 40 .”
ViT-BEVSeg: A Hierarchical Transformer Network for Monocular Birds-Eye-View Segmentation
Dutta, P., Sistu, G., Yogamani, S. K., Galván, E., & McDonald, J. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9891987 · Cited by 19
“We reconfigure the vision transformer [5] in a similar manner to [22] to extract a hierarchical embedded representation of the input image and generate assembled feature maps at different resolutions (as described in Section III).”
MethodAirFormer: Learning-Based Object Detection for Mars Helicopter
Qi, Y., Xiao, X., Yao, M., Xiong, Y., Zhang, L., & Cui, H. (2024) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2024.3492346 · Cited by 25
Realistic Bokeh Effect Rendering on Mobile GPUs, Mobile AI & AIM 2022 Challenge: Report
Ignatov, A., Timofte, R., Zhang, J., Zhang, F., Yu, G., Ma, Z., Wang, H., Kwon, M., Qian, H., Tong, W., et al. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-25066-8_7 · Cited by 17
Monocular Depth Estimation with Adaptive Geometric Attention
Naderi, T., Sadovnik, A., Hayward, J. P., & Qi, H. (2022) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv51458.2022.00069 · Cited by 18
“To mitigate the effect of these convolutions, some researchers have suggested to replace the building blocks entirely or in some places in networks by attention-based blocks [29, 34] or transformers [52, 39] which are themselves, based on attention mechanisms.”
BackgroundRCDPT: Radar-Camera Fusion Dense Prediction Transformer
Lo, C.-C., & Vandewalle, P. (2023) · ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp49357.2023.10096129 · Cited by 18
“The overall structure from DPT [1] is utilized, and multi-modal reassemble layers is proposed.”
BackgroundMethodInfluentialGiezendanner, J., Mukherjee, R., Purri, M., Thomas, M., Mauerman, M., Islam, A. K. M. S., & Tellman, B. (2023) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00209 · Cited by 23
“Additionally, transformers are very data hungry [50], especially when training them on a problem that is very different from the problem they were designed for [34], which makes their use unreasonable for this proof of concept paper.”
BackgroundMonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models
Liu, Y., Fan, K., Yu, W., Li, C., Lü, H., & Yuan, Y. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02009 · Cited by 17
“Then, to effectively consolidate these multi-scale features { F 1 i , ..., F Si } into a unified representation, we adopt the DPT architecture [29], which progressively fuses the feature hierarchy through: where C represents the feature dimension.”
Recurrent Multiscale Feature Modulation for Geometry Consistent Depth Learning
Zhou, Z., Fan, X., Shi, P., Xin, Y., Duan, D., & Yang, L. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3420165 · Cited by 17
“[45]utilizedtheVisionTransformer(ViT) asafoundationalarchitecture,deftlytranslatingitsbag-of-words representationintohierarchical,image-likefeaturemapsforthe assemblyofthefinaldenseprediction.”
Ai, H., Cao, Z., & Wang, L. (2025) · International Journal of Computer Vision · doi:10.1007/s11263-025-02391-w · Cited by 19
“Key directions include, but are not limited to, the collection of large-scale real-world datasets, the development of specialized network architectures to enhance prediction accuracy, and the integration of powerful large-scale vision models to push the boundaries of this field further.”
MPIB: An MPI-Based Bokeh Rendering Framework for Realistic Partial Occlusion Effects
Peng, J., Zhang, J., Luo, X., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20068-7_34 · Cited by 17
Image semantic segmentation of indoor scenes: A survey
Velasteguí, R., Tatarchenko, M., Karaoğlu, S., & Gevers, T. (2024) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2024.104102 · Cited by 23
“…and Transformer-based architectures like SegFormer (Xie et al., 2021), Segmenter (Strudel et al., 2021), SETR (Zheng et al., 2021), and DPT (Ranftl et al., 2021), which adapt the Transformer architecture to image segmentation by utilizing self-attention mechanisms and patch-based…”
Spatiotemporal Vision Transformer for Short Time Weather Forecasting
Bojesomo, A., Al-Marzouqi, H., & Liatsis, P. (2021) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata52589.2021.9671442 · Cited by 20
“The number of blocks in the encoder has been set to four to better handle the data demanding nature of transformers, as the model is not pre-trained on any other dataset [15], [17], [19], [20].”
BackgroundMethodOlabanjo, O. A., Ashiribo, W. U. S. U., Asokere, M., Afisi, O. T., Okugbesan, B., Olabanjo, O., Folorunso, O. O., & Mazzara, M. (2023) · Analytics · doi:10.3390/analytics2030039 · Cited by 21
“(cid:26) Citation The impact index is calculated using Equation (1) below: Impact Index = 2023 − Year Citation otherwise This gives readers an overview of where to publish related research.”
UniMod1K: Towards a More Universal Large-Scale Dataset and Benchmark for Multi-modal Learning
Zhu, X., Xu, T., Liu, Z., Tang, Z., Wu, X., & Kittler, J. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-01999-8 · Cited by 17
“We evaluate several state-of-the-art depth estimation methods on the proposed dataset, including VPD (Zhao et al., 2023b), ZoeDepth (Bhat et al., 2023), NewCRFs (Yuan et al., 2022), MIM (Xie et al., 2023), VADepthNet (Liu et al., 2023), DistDepth (Wu et al., 2022), LapDepth (Song et al., 2021), WaveletMonodepth (Rama-…”
Generative Sparse-View Gaussian Splatting
Kong, H., Yang, X., & Wang, X. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02491 · Cited by 18
“We apply pre-trained DPT [30] to estimate monocular depth for all training and pseudo views.”
InfluentialViDSOD-100: A New Dataset and a Baseline Model for RGB-D Video Salient Object Detection
Lin, J., Zhu, L., Shen, J., Fu, H., Zhang, Q., & Wang, L. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02051-5 · Cited by 23
“Table 3 shows the MAE, F β , and S α results of five state-of-the-art methods (MGA (Li et al, 2019), RCR-Net (Yan et al, 2019), PyramidCSA (Gu et al, et al, 2021) and its advanced iteration, DPT-B+R+AL (Rajpal et al, 2023).”
ThermoPore: Predicting part porosity based on thermal images using deep learning
Pak, P., Ogoke, F., Polonsky, A. T., et al. (2024) · Additive manufacturing · doi:10.1016/j.addma.2024.104503 · Cited by 17
“This network builds off the work by Ranftl et al. 45 where fusion blocks and convolutional layers are added to a vision transformer to provide depth predictions of a given image.”
FAR: Flexible, Accurate and Robust 6DoF Relative Camera Pose Estimation
Rockwell, C., Kulkarni, N. V., Jin, L., Park, J. J., Johnson, J., & Fouhey, D. F. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.01877 · Cited by 17
“+ Solver + Scale”: LoFTR + DPT [57] trained on KITTI [22].”
BackgroundWS-MTST: Weakly Supervised Multi-Label Brain Tumor Segmentation With Transformers
Chen, H., An, J., Jiang, B., Xia, L., Bai, Y., & Gao, Z. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3321602 · Cited by 21
“And Ranftl et al. [26] adopted MLP as the decoder for clearer boundaries.”
BackgroundOn the Importance of Accurate Geometry Data for Dense 3D Vision Tasks
Jung, H., Ruhkamp, P., Zhai, G., Brasch, N., Li, Y., Verdié, Y., Song, J., Zhou, Y., Armagan, A., Ilić, S., et al. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00082 · Cited by 17
MonoVAN: Visual Attention for Self-Supervised Monocular Depth Estimation
Indyk, I., & Makarov, I. (2023) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar59233.2023.00138 · Cited by 16
“…in this area [29,36,37,42] and further improvements in speed and resolution [40,41,44], in particular relying on a transformer-based solution [39,49], utilizing deep ordinal regression network [13], mixing multiple diverse datasets during training [50], and moving from the regression to the…”
Xing, D., Shen, J., Ho, C., & Tzes, A. P. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i3.25401 · Cited by 18
Papa, L., Alati, E., Russo, P., & Amerini, I. (2022) · IEEE Access · doi:10.1109/access.2022.3170425 · Cited by 17
“In Ranftl et al. [20], the authors propose a novel architecture called Dense Prediction Transformer, which uses a transformer encoder combined with a traditional decoder to perform the estimation.”
MethodSelf-Supervised Multi-Frame Monocular Depth Estimation for Dynamic Scenes
Wu, G., Liu, H., Wang, L., Li, K., Guo, Y., & Chen, Z. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3340948 · Cited by 18
“Ranftl et al. [59] achieve fine-grained and globally coherent dense depth estimation by using vision transformers as the backbone.”
BackgroundLDA-Mono: A lightweight dual aggregation network for self-supervised monocular depth estimation
Zhao, B., He, H., Xu, H., Shi, P., Hao, X., & Huang, G. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112552 · Cited by 16
Depth-Cooperated Trimodal Network for Video Salient Object Detection
Lu, Y., Min, D., Fu, K., & Zhao, Q. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897762 · Cited by 19
“To study this, we replace it with Monodepth2 [34] and MegaDepth [35], whose depth prediction accuracy is inferior to DPT, and models are re-trained likewise.”
MethodFacial Animation with Disentangled Identity and Motion using Transformers
Chandran, P., Zoss, G., Groß, M. H., Gotardo, P. F. U., & Bradley, D. (2022) · Computer Graphics Forum · doi:10.1111/cgf.14641 · Cited by 19
Lightweight Monocular Depth with a Novel Neural Architecture Search Method
Huynh, L., Nguyen-Ha, P., Matas, J., Rahtu, E., & Heikkilä, J. (2022) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv51458.2022.00040 · Cited by 19
“Gonzalez and Kim [20] estimated depth by synthesizing stereo pairs from a single image, while [57] and [45] applied vision-transformer for depth prediction.”
BackgroundKick Back & Relax: Learning to Reconstruct the World by Watching SlowTV
Spencer, J., Hadfield, S., Russell, C., & Bowden, R. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01445 · Cited by 19
“The feature-matching baseline [2] consists of LoFTR [58] correspondences, a PnP solver and DPT [46] fine-tuned on either Kitti or NYUD-v2.”
BackgroundMethodInfluentialAdaptive Surface Normal Constraint for Geometric Estimation From Monocular Images
Long, X., Zheng, Y., Zheng, Y., Tian, B., Lin, C. C., Liu, L., Zhao, H., Zhou, G., & Wang, W. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3381710 · Cited by 23
“We use our prior work Cerberus [63], which is inherited from Dense Prediction Trans-former (DPT) [50], as the backbone.”
MethodHeterogeneous Experts and Hierarchical Perception for Underwater Salient Object Detection
Zha, M., Wang, G., Pei, Y., Li, T., Tang, X., Li, C., Yang, Y., & Shen, H. T. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3572760 · Cited by 22
“Due to the lack of corresponding depth maps, we utilize a monocular depth estimation method, i.e., DPT [63], for generation.”
YORO - Lightweight End to End Visual Grounding
Ho, C.-H., Appalaraju, S., Jasani, B. A., Manmatha, R., & Vasconcelos, N. M. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-25085-9_1 · Cited by 19
Monocular Depth Estimation: a Review of the 2022 State of the Art
Ehret, T. (2023) · Image Processing On Line · doi:10.5201/ipol.2023.459 · Cited by 20
“In this paper, we compare five recent deep-learning based monocular depth estimation methods: GPLDepth [15] (2022), Adabins [1] (2021), 3DShape [34] (2021), MiDaS [22] (2020) and DPT [21] (2021).”
BackgroundMethodInfluentialWu, Z., Gan, Y., Li, X., Wu, Y., Wang, X., Xu, T., & Wang, F. (2022) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2022.3218594 · Cited by 16
“Furthermore, the recent works [68], [69] apply vision transformers to improve depth estimation.”
BackgroundRadarCam-Depth: Radar-Camera Fusion for Depth Estimation with Learned Metric Scale
Li, H., Ma, Y., Gu, Y., Hu, K., Liu, Y., & Zuo, X. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10610929 · Cited by 23
Embedding-Free Transformer with Inference Spatial Reduction for Efficient Semantic Segmentation
Yu, H., Cho, Y., Kang, B., Moon, S., Kong, K., & Kang, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72946-1_6 · Cited by 23
Balaha, H. M., Hassan, A. E.-S., Ahmed, R. A., & Balaha, M. H. (2025) · Soft Computing · doi:10.1007/s00500-025-10463-9 · Cited by 19
“Originally grounded in the transformer architecture designed for natural language processing, ViT seamlessly translates its capabilities to the realm of image analysis (Ranftl et al. 2021).”
Conversion of 2D picture to color 3D holography using end-to-end convolutional neural network
Chang, C., Zhao, C., Dai, B., et al. (2025) · PhotoniX · doi:10.1186/s43074-025-00186-3 · Cited by 15
“3, our approach uses an off-the-shelf monocular depth estimation network [40], to firstly predict a depth map for each natural picture.”
InfluentialDG-Recon: Depth-Guided Neural 3D Scene Reconstruction
Ju, J., Tseng, C.-W., Bailo, O., Dikov, G., & Ghafoorian, M. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01667 · Cited by 16
“For a better trade-off between accuracy, feasibility, and computational cost, we adopt the monocular depth estimation approach with a network consisting of a ResNet34 [21] encoder and the convolutional decoder from [32] as proposed in [49].”
BackgroundMethodYang, Q., Gu, J., Xiong, T., et al. (2025) · Agriculture · doi:10.3390/agriculture15181982 · Cited by 15
Yang, P., Liu, H., Roznere, M., et al. (2023) · Springer proceedings in advanced robotics · doi:10.1007/978-3-031-25555-7_7 · Cited by 15
LaVin-DiT: Large Vision Diffusion Transformer
Wang, Z., Xia, X., Chen, R., Yu, D., Wang, C., Gong, M., & Liu, T. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01868 · Cited by 35
“For depth estimation, our model achieves an AbsRel of 6.2 and a threshold accuracy δ 1 of 96.1%, demonstrating competitive performance compared to expert models such as Marigold [28] and DPT [45].”
InfluentialTHE Benchmark: Transferable Representation Learning for Monocular Height Estimation
Xiong, Z., Huang, W. X., Hu, J., & Zhu, X. X. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3311764 · Cited by 29
“Adabins [59] and DenseViT [33] are the state-of-the-art MDE models selected for performance comparison on the proposed GTAH dataset.”
BackgroundMethodPRSeg: A Lightweight Patch Rotate MLP Decoder for Semantic Segmentation
Ma, Y., Lin, F., Wu, S., Tian, S., & Yu, L. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3271523 · Cited by 15
“Some works [26], [41], [64] use top-down connectivity to fuse multi-scale spatial and semantic information to enhance pixel representation and multi-scale object recognition.”
BackgroundSlaBins: Fisheye Depth Estimation using Slanted Bins on Road Environments
Lee, J.-S., Cho, G., Park, J., Kim, K., Lee, S., Kim, J.-H., Jeong, S.-G., & Joo, K. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00805 · Cited by 16
“With the advent of break-throughs in deep neural networks, such as CNN and trans-former architecture, monocular depth estimation has shown impressive performance [15, 33, 21, 32, 23, 18, 40, 11, 9, 12, 29, 19, 16, 17, 20].”
BackgroundMonocular Visual-Inertial Depth Estimation
Wofk, D., Ranftl, R., Müller, M., & Koltun, V. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161013 · Cited by 18
“Let z refer to unit-less affine-invariant inverse depth that is output by a monocular depth estimation model such as DPT-Hybrid.”
InfluentialSelf-citationDeep Learning for Omnidirectional Vision: A Survey and New PerspectivesPreprint
Ai, H., Cao, Z., Zhu, J., Bai, H., Chen, Y., & Wang, L. (2022) · doi:10.36227/techrxiv.19807699.v1 · Cited by 15
Lu, Y., Wei, R., Li, B., Chen, W., Zhou, J., Dou, Q., Sun, D. B., & Liu, Y. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10161505 · Cited by 18
“we build the decoder [15], each layer of which for depth estimation consists of three parts:”
BackgroundMethodLoopSparseGS: Loop-Based Sparse-View Friendly Gaussian Splatting
Bao, Z., Liao, G., Zhou, K., Liu, K., Li, Q., & Qiu, G. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3574929 · Cited by 15
“Our approach is implemented using the PyTorch framework, and utilizes the pre-trained Midas Model [28, 29] for zero-shot monocular depth estimation and the Colmap model for initial camera poses and 3D points.”
MethodLearning Phase Mask for Privacy-Preserving Passive Depth Estimation
Tasneem, Z., Milione, G., Tsai, Y., Yu, X., Veeraraghavan, A., Chandraker, M. K., & Pittaluga, F. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20071-7_30 · Cited by 17
“For all sensors, we show the corresponding face identification accuracy of E � cientNet-b0 [50] on the VGGFace2 dataset, along the x-axis, and the depth estimation accuracy ( � 1 ) of DPT [43] on the NYU Depth v2 dataset, along the y-axis.”
MethodOASL: Orientation-aware adaptive sampling learning for arbitrary oriented object detection
Zhao, Z., & Li, S. (2024) · International Journal of Applied Earth Observation and Geoinformation · doi:10.1016/j.jag.2024.103740 · Cited by 15
Readout Guidance: Learning Control from Diffusion Features
Luo, G., Darrell, T., Wang, O., Goldman, D. B., & Holynski, A. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.00785 · Cited by 35
“We use a similar setup to ControlNet [70], in which pseudo labels for the control are computed by off-the-shelf models and standardized into RGB images; we use OpenPose [8] to extract human pose, MiDaS [48, 49] for depth, and HED [65] for edges.”
MethodMake-It-4D: Synthesizing a Consistent Long-Term Dynamic Scene Video from a Single Image
Shen, L., Li, X., Sun, H., Peng, J., Xian, K., Cao, Z., & Lin, G. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3612033 · Cited by 17
“After that, we employ a pre-trained monocular depth estimator [26] to predict the depth map of the outpainted image and leverage depth discontinuities to partition the outpainted image into layered depth images (LDIs) [24, 32].”
BackgroundMethodDAUT: Underwater Image Enhancement Using Depth Aware U-shape Transformer
Badran, M., & Torki, M. (2023) · Proceedings - International Conference on Image Processing · doi:10.1109/icip49359.2023.10222895 · Cited by 14
“The depth estimation stage used in our network is the dense vision transformer (DPT) [10].”
MethodZeST: Zero-Shot Material Transfer from a Single Image
Cheng, T.-Y., Sharma, P., Markham, A., Trigoni, N., & Jampani, V. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73232-4_21 · Cited by 28
Natan, O., & Miura, J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-02375-0_26 · Cited by 14
The Third Monocular Depth Estimation Challenge
Spencer, J., Tosi, F., Poggi, M., Arora, R. S., Russell, C., Hadfield, S., Bowden, R., Zhou, G., Li, Z., Rao, Q., et al. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00005 · Cited by 21
“Deeper network architectures [80, 109], multi-scale fusion [63], and transformer-based encoders [8, 16, 79] further enhanced performance.”
BackgroundMethodPeng, C., & Zhang, J. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00313 · Cited by 15
“2018 [38], MegaDepth [15], MiDaS v2 [21] and v3 [22], SGR [39], Huynh et al.”
BackgroundExpediting Large-Scale Vision Transformer for Dense Prediction Without Fine-Tuning
Yuan, Y., Liang, W., Ding, H., Liang, Z., Zhang, C., & Hu, H. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3327511 · Cited by 14
“Illustrating the improvements of our approach: we report the results of applying our approach to Segmenter [28] for semantic segmentation and DPT [29] for depth estimation on the 1-st and 2-ed row respectively.”
InfluentialMAMo: Leveraging Memory and Attention for Monocular Video Depth Estimation
Yasarla, R., Cai, H., Jeong, J., Shi, Y., Garrepalli, R., & Porikli, F. (2023) · 2023 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00804 · Cited by 15
“By leveraging deep learning, researchers have achieved significantly more accurate image-based depth estimation [11, 13, 3, 43, 69].”
BackgroundMethodInfluentialEgoLoc: Revisiting 3D Object Localization from Egocentric Videos with Visual Queries
Mai, J., Hamdi, A., Giancola, S., Zhao, C., & Ghanem, B. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00011 · Cited by 23
“We adopt the DPT [67] network pretrained on NYU V2 [60] for depth estimation.”
MethodJaramillo-Hernández, J. F., Julián, V., Marco-Detchart, C., & Rincon, J. A. (2024) · Sensors · doi:10.3390/s24030937 · Cited by 13
“This requires incorporating representative depth labels for each object using the state-of-the-art monocular depth estimation model MiDaS [31,32].”
BackgroundMethodOpen-Vocabulary Camouflaged Object Segmentation
Pang, Y., Zhao, X., Zuo, J., Zhang, L., & Lu, H. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72970-6_27 · Cited by 18
Learned Monocular Depth Priors in Visual-Inertial Initialization
Zhou, Y., Kar, A., Turner, E., Kowdle, A., Guo, C. X., DuToit, R. C., & Tsotsos, K. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20047-2_32 · Cited by 15
S2Gaussian: Sparse-View Super-Resolution 3D Gaussian Splatting
Wan, Y., Shao, M., Cheng, Y., & Zuo, W. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00075 · Cited by 15
“For the optimization of LR GS, the pre-trained DPT [28] model is adapted for depth estimation and the total iterations are set to 10,000.”
Anatomically accurate cardiac segmentation using Dense Associative Networks
Ullah, Z., & Kim, J. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.112742 · Cited by 22
RoI-Attention Network for Small Disease Segmentation in Crop Images
Moon, G.-Y., & Kim, J. (2024) · IEEE Access · doi:10.1109/access.2024.3393301 · Cited by 14
“For instance, vision transformer based DPT [28] requires a significant amount of data for effective training.”
BackgroundResultRankSeg: Adaptive Pixel Classification with Image Category Ranking for Segmentation
He, H., Yuan, Y., Yue, X., & Hu, H. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19818-2_39 · Cited by 16
“, high-resolution representation learning [60,70], contextual representation aggregation [12,23,35,87,81], boundary refinement [66,43,84], and vision transformer architecture designs [53,61,82,88].”
BackgroundDepth-Guided Sparse Structure-from-Motion for Movies and TV Shows
Liu, S., Nie, X., & Hamid, R. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01551 · Cited by 19
“This is because the estimated depth images [30] of initial image-pair for this example are too erroneous for our approach to effectively guide the subsequent reconstruction process to a better solution.”
InfluentialDepth-aware guidance with self-estimated depth representations of diffusion models
Kim, G., Jang, W., Lee, G., Hong, S., Seo, J., & Kim, S. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110474 · Cited by 33
SENSE: Self-Evolving Learning for Self-Supervised Monocular Depth Estimation
Li, G., Huang, R., Li, H., You, Z., & Chen, W. (2023) · IEEE Transactions on Image Processing · doi:10.1109/tip.2023.3338053 · Cited by 15
“As shown by works [8], [9], and [10], feature resolution is important for dense prediction tasks, like semantic segmentation and depth prediction.”
BackgroundMSM-ViT: A multi-scale MobileViT for pulmonary nodule classification using CT images
Cao, K., Tao, H., Wang, Z., et al. (2023) · Journal of X-Ray Science and Technology · doi:10.3233/xst-230014 · Cited by 14
LAM: Large Avatar Model for One-shot Animatable Gaussian Head
He, Y., Gu, X., Ye, X. W., Xu, C., Zhao, Z., Dong, Y., Yuan, W., Dong, Z., & Bo, L. (2025) · ACM SIGGRAPH Conference (SIGGRAPH) · doi:10.1145/3721238.3730706 · Cited by 17
“To address this issue, we incorporate insights from the animatable FLAME model, which is animatable without extra neural networks.”
Measuring and Modeling Uncertainty Degree for Monocular Depth Estimation
Xiang, M., Zhang, J., Barnes, N. M., & Dai, Y. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3362357 · Cited by 15
Drone-Person Tracking in Uniform Appearance Crowd: A New Dataset
Alansari, M., Hay, O. A., Alansari, S., Javed, S., Shoufan, A., Zweiri, Y., & Werghi, N. (2024) · Scientific Data · doi:10.1038/s41597-023-02810-y · Cited by 13
$\mathcal {S}^{2}$Net: Accurate Panorama Depth Estimation on Spherical Surface
Li, M., Wang, S., Yuan, W., Shen, W., Sheng, Z., & Dong, Z. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3234820 · Cited by 13
“The Base fusion module is only consisted of two simple residual convolution units from work [5] using basic spherical convolution method in previous works [14], [25].”
BackgroundMethodInfluentialMonocular Depth Estimation for Glass Walls With Context: A New Dataset and Method
Liang, Y., Deng, B., Liu, W., Qin, J., & He, S. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3308551 · Cited by 17
“To demonstrate the performance of traditional depth estimation methods [8], [16], [17], [30], [56], [57], [58] on transparent glass walls, we evaluate the models trained on NYU depth V2 [9] and tested on our GW-Depth dataset.”
Method
Metadata
- Authors
- Ranftl, R., Bochkovskiy, A., & Koltun, V.
- Affiliations
- Intel Labs
- Intel (United States) (US)
- Venue
- 2021 IEEE/CVF International Conference on Computer Vision (ICCV)
- Publisher
- IEEE
- Published
- 2021-10
- Pages
- 12159-12168
- OpenAlex
- W3136635488
- Cited by
- 2,373
- References
- 88
- Citations by year
- 2026: 480 · 2025: 607 · 2024: 586 · 2023: 431 · 2022: 225 · 2021: 44
- FWCI
- 104.94
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE/CVF International Conference on Computer Vision (ICCV) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- TIB Data Manager
- IEEE/CVF International Conference on Computer Vision (ICCV)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Keywords
- vision transformer, dense prediction, monocular depth estimation, semantic segmentation, convolutional decoder, global receptive field, Pascal Context, NYUv2, KITTI, ADE20K
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
