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 201–300 of 2,248
A semi‐automated camera trap distance sampling approach for population density estimation
Henrich, M., Burgueño, M., Hoyer, J., Haucke, T., Steinhage, V., Kühl, H. S., & Heurich, M. (2023) · Remote Sensing in Ecology and Conservation · doi:10.1002/rse2.362 · Cited by 44
Physically-Based Editing of Indoor Scene Lighting from a Single Image
Li, Z., Shi, J. D., Bi, S., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20068-7_32 · Cited by 42
Learning Graph Regularisation for Guided Super-Resolution
de Lutio, R., Becker, A., D’Aronco, S., Russo, S., Wegner, J. D., & Konrad Schindler, B. (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/cvpr52688.2022.00202 · Cited by 53
“Finally, to explore the boundaries of feature extraction, we employ a variant of the Dense Prediction Transformer [38, 39], which extracts multi-scale features for dense prediction tasks (like monodepth or semantic segmentation) from a Vision Transformer (ViT).”
MethodLuo, M., Zhang, T., Wei, S., & Ji, S. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3460085 · Cited by 53
“Restrictions apply. its shortcomings of only generating 1/16-scale features for small object detection in remote sensing images, we introduce an adapter inspired by ViT-Adapter [58], as shown in Fig.”
MethodJiang, M., Ma, J., Chen, J., Wang, Y., & Fang, X. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111597 · Cited by 39
Object Segmentation by Mining Cross-Modal Semantics
Wu, Z., Wang, J., Zhou, Z., et al. (2023) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3581783.3611970 · Cited by 41
“Following previous works [64, 68], we generate pseudo-depth from the input RGB image [40, 41].”
MethodMETER: A Mobile Vision Transformer Architecture for Monocular Depth Estimation
Papa, L., Russo, P., & Amerini, I. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3260310 · Cited by 43
“Ranftl et al. [4] investigate the application of ViT proposing DPT, a model composed of a transformer-CNN encoder and a fully-convolutional decoder.”
BackgroundMethodMulti-scale hybrid vision transformer and Sinkhorn tokenizer for sewer defect classification
Haurum, J. B., Madadi, M., Escalera, S., & Moeslund, T. B. (2022) · Automation in Construction · doi:10.1016/j.autcon.2022.104614 · Cited by 46
Consistent-1-to-3: Consistent Image to 3D View Synthesis via Geometry-aware Diffusion Models
Ye, J., Wang, P., Li, K., Shi, Y., & Wang, H. (2024) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv62453.2024.00027 · Cited by 78
“For example, transformer models for single view reconstruction, e.g. DPT [31], GANs [10] and 2D diffusion models for image generation [7].”
MethodZhou, W., Tang, B., Cong, R., & Jiang, Q. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3648880 · Cited by 47
“We used the DPT [69] in USOD10K to generate corresponding depth maps for the USOD dataset, which is exclusively for testing purposes.”
BRNet: Exploring Comprehensive Features for Monocular Depth Estimation
Han, W., Yin, J., Jin, X., Xiangdong, D., & Shen, J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19839-7_34 · Cited by 49
“And then, many methods are exploited to improve the performance [2,31,25].”
MethodUnderstanding the brain with attention: A survey of transformers in brain sciences
Chen, C., Wang, H., Chen, Y., Yin, Z., Yang, X., Ning, H., Zhang, Q., Li, W., Xiao, R., & Zhao, J. (2023) · Brain‐X · doi:10.1002/brx2.29 · Cited by 44
Neural Canvas: Supporting Scenic Design Prototyping by Integrating 3D Sketching and Generative AI
Shen, Y., Shen, Y., Cheng, J., Jiang, C., Fan, M., & Wang, Z. (2024) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3613904.3642096 · Cited by 46
“One potential solution is using image depth estimation algorithms [56, 66] to automatically provide users with candidate layers.”
Point Cloud Forecasting as a Proxy for 4D Occupancy Forecasting
Khurana, T., Hu, P., Held, D., & Ramanan, D. (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/cvpr52729.2023.00114 · Cited by 40
“Depth on sky regions is untrustworthy as no returns are received for this region from the LiDAR sensor. makes it easy to train across different datasets, making zero-shot cross-dataset transfer possible for LiDARs [14,15].”
BackgroundEvent-Based Monocular Depth Estimation With Recurrent Transformers
Liu, X., Li, J., Shi, J., Xiaopeng Fan, B., Tian, Y., & Zhao, D. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3378742 · Cited by 52
“More recently, transformers [26], [27], [28] demonstrate appealing potential in modeling global spatial context information for frame-based monocular depth estimation tasks [21], [29], [30], [31].”
BackgroundMethodInfluentialT2I-CompBench++: An Enhanced and Comprehensive Benchmark for Compositional Text-to-Image Generation
Huang, K., Duan, C., Sun, K., Xie, E., Li, Z., & Liu, X. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3531907 · Cited by 195
“We compute the overall score by multiplying the probability of answering “yes” for each question.”
Attention Guided CAM: Visual Explanations of Vision Transformer Guided by Self-Attention
Leem, S., & Seo, H. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i4.28077 · Cited by 48
Language-driven Grasp Detection
Vuong, A. D., Vu, M. N., Huang, B., Nguyen, N., Le, H., Vo, T., & Nguyen, A. (2024) · 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52733.2024.01695 · Cited by 39
“First, future work could investigate the use of text or image-to-3D models [95] or image-to-depth [73] and reuse our dataset’s prompts and images to construct 3D language-driven grasp datasets.”
BackgroundHuang, Y., Kang, D., Jia, W., Liu, L., & He, X. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i1.19985 · Cited by 45
“CTNet (Li, Sun, and Tang 2021) + JPU 55.5 SETR-MLA (Sixiao et al. 2021) 55.83 HRNetV2 + OCR (Ding et al. 2019) 56.2 ResNeSt-269 (Zhang et al. 2020) + DeepLab V3+ 58.9 HRNetV2 + OCR + RMI 59.6 DPT-Hybrid (Ranftl, Bochkovskiy, and Koltun 2021) 60.46 Our CAA (EfficientNet-B7, w/o decoder) 60.12 Our CAA (EfficientNet-B7 + s…”
InfluentialAodeMar: Attention-Aware Occlusion Detection of Vessels for Maritime Autonomous Surface Ships
Wang, N., Wang, Y., Feng, Y., & Wei, Y. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3398733 · Cited by 42
“Inspired by successes in NLP, multiple SAMs including vision transformer [35], [68] and STR [36] have been extensively employed to capture features that appear simultaneously in the same scene.”
BackgroundMethodScaled ReLU Matters for Training Vision Transformers
Wang, P., Wang, X., Luo, H., et al. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i3.20150 · Cited by 38
“…by dividing the images into non-overlapping patches, many ViT variants are pro-a posed (Wang et al. 2021b; Han et al. 2021; Chen et al. 2021a; Ranftl, Bochkovskiy, and Koltun 2021; Liu et al. 2021; Chen, Fan, and Panda 2021; Zhang et al. 2021a; Xie et al. 2021; Zhang et al. 2021b;…”
BackgroundPrompt Guided Transformer for Multi-Task Dense Prediction
Lu, Y., Sirejiding, S., Ding, Y., Wang, C., & Lu, H. (2024) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2024.3349865 · Cited by 37
“ViT-based models demonstrate remarkable performance on various vision tasks, such as image classification [9], [18], [22], [26], object detection [5], [66], dense prediction [7], [36], [40], [49], and video understanding [13], [50], surpassing previous CNN-based models.”
MethodWang, Y., Shi, M., Li, J., Huang, Z., Cao, Z., Zhang, J., Xian, K., & Lin, G. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00868 · Cited by 39
Phaseformer: Phase-Based Attention Mechanism for Underwater Image Restoration and Beyond
Khan, M. R., Negi, A., Kulkarni, A., et al. (2025) · IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv61041.2025.00931 · Cited by 38
Transformer Scale Gate for Semantic Segmentation
Shi, H., Hayat, M., & Cai, J. (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/cvpr52729.2023.00298 · Cited by 42
Deep Learning Model for Real‑time Semantic Segmentation During Intraoperative Robotic Prostatectomy
Park, S. G., Park, J., Choi, H. R., Lee, J. H., Cho, S. T., Lee, Y. G., Ahn, H., & Pak, S. (2024) · European Urology Open Science · doi:10.1016/j.euros.2024.02.005 · Cited by 46
GEDepth: Ground Embedding for Monocular Depth Estimation
Yang, X., Ma, Z., Ji, Z., & Ren, Z. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01168 · Cited by 46
Recent Advances in Conventional and Deep Learning-Based Depth Completion: A Survey
Xie, Z., Yu, X., Gao, X., Li, K., & Shen, S. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3201534 · Cited by 37
“…applied in the national language processing (NLP) [166] community, and the computer vision (CV) [167] community more recently, for example, for object detection [168], high-resolution image synthesis [169], point cloud completion [170], depth estimation and semantic segmentation [171], and so on.”
BackgroundHuang, Y., Sanatani, R. P., Liu, C., et al. (2024) · Building and Environment · doi:10.1016/j.buildenv.2024.112395 · Cited by 37
Iqbal, S., Khan, T. M., Naqvi, S. S., Naveed, A., Usman, M., Khan, H. A., & Razzak, I. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3331278 · Cited by 38
“Severalsubsequenttechniqueshavebeendevelopedthatexpand onViTs,andmanysemanticsegmentationalgorithms[62],[63], [64],[65]basedonViTbackboneshaveachievedpromisingre-sultsduetothepowerfulrepresentationslearnedfrompretrained backbones.”
EGformer: Equirectangular Geometry-biased Transformer for 360 Depth Estimation
Yun, I., Shin, C., Lee, H., Lee, H.-J., & Rhee, C.-E. (2023) · IEEE International Conference on Computer Vision · doi:10.1109/iccv51070.2023.00561 · Cited by 69
“Compared to a CNN, ViT [14] possesses a global receptive field, which is highly beneficial for various vision tasks [6, 17, 52, 47, 28, 21].”
BackgroundMethodDeep Learning-Driven Automatic Segmentation of Weeds and Crops in UAV Imagery
Tao, J., Qiao, Q., Song, J., Sun, S., Chen, Y., Wu, Q., Liu, Y., Xue, F., Wu, H., & Zhao, F. (2025) · Sensors · doi:10.3390/s25216576 · Cited by 40
AffordanceLLM: Grounding Affordance from Vision Language Models
Qian, S., Chen, W., Bai, M., Zhou, X. H., Tu, Z., & Li, L. E. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00754 · Cited by 59
“For each image, we use the state-of-the-art depth estimation model DPT [55] to generate a pseudo depth map D .”
BackgroundMethodInfluentialPastNet: Introducing Physical Inductive Biases for Spatio-temporal Video Prediction
Wu, H., Xu, F., Chen, C., Hua, X., Luo, X., & Wang, H. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3681489 · Cited by 43
“As a dense prediction problem [36], the scalability of neural network models is critical for high-resolution video prediction [5].”
BackgroundFlash3D: Feed-Forward Generalisable 3D Scene Reconstruction from a Single Image
Szymanowicz, S., Insafutdinov, E., Zheng, C., Campbell, D., Henriques, J. F., Rupprecht, C., & Vedaldi, A. (2025) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv66043.2025.00067 · Cited by 39
“Approaches like Viewset Diffusion [71] and RenderDiffusion [2] fuse 3D reconstruction with diffusion-based generation, which can reduce but not eliminate the cost of iterative generation.”
CATNet: Convolutional attention and transformer for monocular depth estimation
Tang, S., Lu, T., Liu, X., Zhou, H., & Zhang, Y. (2023) · Pattern Recognition · doi:10.1016/j.patcog.2023.109982 · Cited by 38
Enlarged Motion-Aware and Frequency-Aware Network for Compressed Video Artifact Reduction
Liu, W., Gao, W., Li, G., et al. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3406425 · Cited by 36
“In particular, multi-frame enhancement methods can be divided into methods based on explicit or implicit optical flow estimation (OFE) [7], [8], [9], [20], [21], [22], methods based on deformable convolution (DCN) [10], [11], [23], [24], [25] and methods based on self attention [26].”
BackgroundMethodContent-aware Token Sharing for Efficient Semantic Segmentation with Vision Transformers
Lu, C., de Geus, D., & Dubbelman, G. (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/cvpr52729.2023.02263 · Cited by 36
Li, R., Ji, P., Xu, Y., & Bhanu, B. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3207105 · Cited by 35
“Ranftl et al. [35] propose robust scale-and shift-invariant losses for training the model using data from mixed dataset and testing on zero-shot datasets, and improve it further by integrating vision transformer in network design [36].”
MethodEdgeViT: Efficient Visual Modeling for Edge Computing
Chen, Z., Zhong, F., Luo, Q., Zhang, X., & Zheng, Y. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19211-1_33 · Cited by 44
Towards Training-Free Open-World Segmentation via Image Prompt Foundation Models
Tang, L., Jiang, P.-T., Xiao, H., & Li, B. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02185-6 · Cited by 38
“These methods (Luo et al, 2023; Xu et al, 2023; Ma et al, 2022; Liang et al, 2023; Xu et al, 2022; Zhou et al, 2023b; Liu et al, 2022) adapt CLIP models to the open-world segmentation models by training on the datasets with seen classes to align the predicted region features and text features.”
Generative Transformer for Accurate and Reliable Salient Object Detection
Mao, Y., Zhang, J., Wan, Z., Tian, X., Li, A., Lv, Y., & Dai, Y. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3469286 · Cited by 96
“3) Different Transformer Backbones Analysis: Following our pipeline, we change the Swin transformer backbone [21] to the ViT backbone [20], [30], and achieve TIGAN_ViT in Table IV.”
InfluentialAnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views
Jiang, L., Mao, Y., Xu, L., Lu, T., Ren, K., Jin, Y., Xu, X., Yu, M., Pang, J., Zhao, F., et al. (2025) · ACM Transactions on Graphics · doi:10.1145/3763326 · Cited by 46
“2, we adopt a dual-head design based on the DPT decoder [29] to predict all Gaussian parameters.”
EndoViT: pretraining vision transformers on a large collection of endoscopic images
Batić, D., Holm, F., Özsoy, E., Czempiel, T., & Navab, N. (2024) · International Journal of Computer Assisted Radiology and Surgery · doi:10.1007/s11548-024-03091-5 · Cited by 51
“Since the ViT backbone processes the input at high resolution and has a global receptive field, DPT allows for fine-grained and more globally consistent predictions compared to previous CNN approaches, especially when a larger amount of data can be provided [22].”
BackgroundMethodDoF-NeRF: Depth-of-Field Meets Neural Radiance Fields
Wu, Z., Li, X., Peng, J., Lü, H., Cao, Z., & Zhong, W. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548088 · Cited by 34
“We extend Table 2 [27] and the predicted disparity map from DPT [32].”
MethodFiFoNet: Fine-Grained Target Focusing Network for Object Detection in UAV Images
Xi, Y., Jia, W., Miao, Q., Liu, X., Fan, X., & Li, H. (2022) · Remote Sensing · doi:10.3390/rs14163919 · Cited by 35
Transformers for mapping burned areas in Brazilian Pantanal and Amazon with PlanetScope imagery
Gonçalves, D. N., Marcato, J., Carrilho, A. C., et al. (2022) · International Journal of Applied Earth Observation and Geoinformation · doi:10.1016/j.jag.2022.103151 · Cited by 34
Exploring Efficiency of Vision Transformers for Self-Supervised Monocular Depth Estimation
Karpov, A., & Makarov, I. (2022) · Proceedings - International Symposium on Mixed and Augmented Reality, ISMAR · doi:10.1109/ismar55827.2022.00089 · Cited by 36
“Recent studies [3, 38, 53] introduced vision transformer for depth estimation.”
BackgroundMethodInfluentialFraudGT: A Simple, Effective, and Efficient Graph Transformer for Financial Fraud Detection
Lin, J., Guo, X., Zhu, Y., Mitchell, S., Altman, E., & Shun, J. (2024) · ACM International Conference on AI in Finance (ICAIF) · doi:10.1145/3677052.3698648 · Cited by 36
Interaction Transformer for Human Reaction Generation
Chopin, B., Tang, H., Otberdout, N., Daoudi, M., & Sebe, N. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3242152 · Cited by 49
“They also have recently demonstrated good performance on a broad range of tasks such as image classification [36], image generation [37], [38], object detection [39], [40], human pose estimation [41], depth estimation [42], [43], 3D pose transfer [44], [45], and action recognition [32], [46].”
BackgroundSun, L. C., Bhatt, N. P., Liu, J. C., Fan, Z., Wang, Z., Humphreys, T. E., & Topcu, U. (2024) · 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802389 · Cited by 34
“However, in the absence of a such sensors, monocular dense depth estimation networks, such as DPT [26], can be used.”
Survey on Monocular Metric Depth Estimation
Zhang, J., Wu, Y., & Jiang, H. (2025) · Computers · doi:10.3390/computers14110502 · Cited by 46
Reference-guided Controllable Inpainting of Neural Radiance Fields
Mirzaei, A., Aumentado‐Armstrong, T., Brubaker, M. A., Kelly, J. W., Levinshtein, A., Derpanis, K. G., & Gilitschenski, I. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01633 · Cited by 38
“To address this challenge, we propose the use of a monocular depth estimator [52, 51], D̃(·), to predict the uncalibrated disparity of the”
MethodXiang, J., Wang, Y., An, L., Liu, H., Wang, Z., & Liu, J. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3210298 · Cited by 38
“Nowadays, attention mechanisms have been applied in many visual tasks [37], [38], [39], [40], [41], [43], [44].”
BackgroundSharingan: A Transformer Architecture for Multi-Person Gaze Following
Tafasca, S., Gupta, A., & Odobez, J. (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/cvpr52733.2024.00196 · Cited by 33
“Our design is inspired by DPT [28], which can only handle decoding image tokens alone.”
MethodVisual SLAM Based on Semantic Segmentation and Geometric Constraints for Dynamic Indoor Environments
Yang, S., Zhao, C., Wu, Z., Wang, Y., Wang, G., & Li, D. (2022) · IEEE Access · doi:10.1109/access.2022.3185766 · Cited by 33
“As a new module, this thread mainly obtains semantic information in the scene through the semantic segmentation network DPT [30].”
DPF-Nutrition: Food Nutrition Estimation via Depth Prediction and Fusion
Han, Y., Cheng, Q., Wu, W., & Huang, Z. (2023) · Foods · doi:10.3390/foods12234293 · Cited by 35
“More detailed implementation of DPT can be obtained in the paper [24].”
BackgroundMethodMUSt3R: Multi-view Network for Stereo 3D Reconstruction
Cabon, Y., Stoffl, L., Antsfeld, L., Csurka, G., Chidlovskii, B., Revaud, J., & Leroy, V. (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/cvpr52734.2025.00106 · Cited by 75
DiffusionLight: Light Probes for Free by Painting a Chrome Ball
Phongthawee, P., Chinchuthakun, W., Sinsunthithet, N., Jampani, V., Raj, A., Khungurn, P., & Suwajanakorn, S. (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/cvpr52733.2024.00018 · Cited by 50
“Then, we paint a circle both at the depth map’s center with the distance closest to the camera (visualized as white) and in an inpaint mask.”
MethodSpike Transformer: Monocular Depth Estimation for Spiking Camera
Zhang, J., Tang, L., Yu, Z., Lu, J., & Huang, T. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20071-7_3 · Cited by 34
“These attractive properties can naturally lead to fine-grained and globally coherent dense predictions [46].”
BackgroundMethodContributions of Shape, Texture, and Color in Visual Recognition
Ge, Y., Xiao, Y., Xu, Z. J., Wang, X., & Itti, L. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19775-8_22 · Cited by 43
Fully Transformer Networks for Semantic Image SegmentationPreprint
Wu, S., Wu, T., Lin, F., Tian, S., & Guo, G. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.04108 · Cited by 64
“DPT [38] assembled tokens from multiple stages of the vision transformer and progressively combined them into full-resolution predictions using a convolutional decoder.”
BackgroundMethodInfluentialUnified Multi-Weather Visibility Restoration
Kulkarni, A., Patil, P. W., Murala, S., & Gupta, S. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3225712 · Cited by 35
“This affects the accuracy of outdoor computer vision algorithms for various tasks such as object detection [1], semantic segmentation [2], scene analysis [3], depth estimation [4], moving object segmentation [5], [6] etc. which are widely used as surveillance applications.”
BackgroundMethodLong-term effects of redlining on climate risk exposure
Miranda, A. S., Conzelmann, C., Phan, T., & Hoffman, J. S. (2024) · Nature Cities · doi:10.1038/s44284-024-00076-y · Cited by 40
FLARE: Feed-Forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views
Zhang, S., Wang, J., Xu, Y., Xue, N., Rupprecht, C., Zhou, X., Shen, Y., & Wetzstein, G. (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/cvpr52734.2025.02043 · Cited by 37
Lima, A. C. D. M., de Paiva, L. F., Bráz, G., et al. (2023) · IEEE Access · doi:10.1109/access.2023.3297097 · Cited by 31
“For this purpose, a pre-trained Dense Prediction Transformer (DPT) model [38] was used.”
MethodCRFormer: Cross-Resolution Transformer for segmentation of grape leaf diseases with context mining
Zhang, X., Cen, C., Li, F., Liu, M., & Mu, W. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.120324 · Cited by 32
Kodipalli, A., Fernandes, S. L., & Dasar, S. K. (2024) · Diagnostics · doi:10.3390/diagnostics14050543 · Cited by 31
“Rene et al. [18] highlight the strength of transformers in semantic segmentation, considering various parameters that can enhance the ability of a transformer in the segmentation process.”
BackgroundGeoRGS: Geometric Regularization for Real-Time Novel View Synthesis From Sparse Inputs
Liu, Z., Su, J., Cai, G., Chen, Y., Zeng, B., & Wang, Z. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3447113 · Cited by 33
“SparseNeRF [20] does not rely on precise depth priors, instead, it supervises training views using depth ranking knowledge [21].”
Learning Depth Estimation for Transparent and Mirror Surfaces
Costanzino, A., Ramirez, P. Z., Poggi, M., Tosi, F., Mattoccia, S., & Stefano, L. D. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00848 · Cited by 38
“When using N = 1 we generate a single in-painted image that is forwarded to the monocular network, while with N = 5 we generate virtual depths from 5 masked images with different colors which are then aggregated by selecting the pixel-wise MiDaS[36] DPT[37] Category depth median.”
BackgroundMethodInfluentialSemantic Segmentation by Early Region Proxy
Zhang, Y., Pang, B., & Lu, C. (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/cvpr52688.2022.00132 · Cited by 35
“…segmentation performance: 1) Enlarge receptive field using larger kernels or dilated convolution [4, 5, 33, 51]; 2) Inte-grate multi-scale features [6,7,22,54]; 3) Refine contextual information [26,52,53]; 4) Utilize attention [15,21,24,57]; 5) Search or design designated backbones [25,32,34,45].”
BackgroundNDDepth: Normal-Distance Assisted Monocular Depth Estimation and Completion
Shao, S., Pei, Z., Chen, W., Chen, P. C. Y., & Li, Z. G. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3411571 · Cited by 33
Overcoming the distance estimation bottleneck in estimating animal abundance with camera traps
Haucke, T., Kühl, H. S., Hoyer, J., & Steinhage, V. (2021) · Ecological Informatics · doi:10.1016/j.ecoinf.2021.101536 · Cited by 31
“The N reference images Irefi are first propagated through the DPT (Ranftl et al., 2021) depth estimation model which results in N uncalibrated disparity images Drefi depicting the pixel-wise inverse distances to scene objects in a relative way, i.e., depth image pixels in blue are closer than those…”
InfluentialYun, I., Lee, H.-J., & Rhee, C. E. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i3.20231 · Cited by 30
“Ours w/ FB represents the model using normal fusion blocks (Ranftl, Bochkovskiy, and Koltun 2021), while Ours w/ NLFB is the model us-ing the proposed non-local fusion blocks.”
MethodNTIRE 2023 Challenge on HR Depth from Images of Specular and Transparent Surfaces
Ramirez, P. Z., Tosi, F., Stefano, L. D., Timofte, R., Costanzino, A., Poggi, M., Salti, S., Mattoccia, S., Shi, J., Zhang, D., et al. (2023) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00143 · Cited by 29
“The first entry on top reports the results by the baseline method – i.e., the DPT [46] model, fine-tuned on the Booster training split as detailed in [76].”
MethodSelf-Supervised Monocular Depth Estimation Using Hybrid Transformer Encoder
Hwang, S.-J., Park, S. J., Baek, J.-H., & Kim, B. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3199265 · Cited by 31
“Transformer is a recently popular backbone network in the field of depth estimation [13], [14].”
BackgroundMethodParida, K. K., Srivastava, S., & Sharma, G. (2022) · 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv51458.2022.00221 · Cited by 29
“Following [26], we obtain features from four different layers of the transformer network, i.”
MethodAn Overview of Vision Transformers for Image Processing: A Survey
Kameswari, C. S., Kavitha, J., Reddy, T. S., Chinthaguntla, B., Jagatheesaperumal, S. K., Gaftandzhieva, S. N., & Doneva, R. Z. (2023) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2023.0140830 · Cited by 38
“The Dense Prediction Transformer (DPT) [80] is a neural network design that successfully uses visual transformers for dense prediction problems.”
Background3D reconstruction from endoscopy images: A survey
Yang, Z., Dai, J. P., & Pan, J. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.108546 · Cited by 35
VCoder: Versatile Vision Encoders for Multimodal Large Language Models
Jain, J., Yang, J., & Shi, H. (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/cvpr52733.2024.02644 · Cited by 53
“For getting depth maps, we use the publicly available ViT-L/14 distilled variant of DINOv2 [52] DPT [57] trained on the NYUd [49] dataset.”
BackgroundMethodInfluentialSelf-distilled Feature Aggregation for Self-supervised Monocular Depth Estimation
Zhou, Z., & Dong, Q. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19769-7_41 · Cited by 36
Bojesomo, A., Al-Marzouqi, H., & Liatsis, P. (2023) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2023.3323729 · Cited by 34
“Recently, self-attention transformer-based models have gradually revolutionized computer vision applications [24], [25], [26], [27].”
BackgroundInfluentialGurses, M. E., González, N., Xu, Y., et al. (2024) · Journal of neurosurgery · doi:10.3171/2023.10.jns23516 · Cited by 29
ComboVerse: Compositional 3D Assets Creation Using Spatially-Aware Diffusion Guidance
Chen, Y., Wang, T., Wu, T., Pan, X., Jia, K., & Liu, Z. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72691-0_8 · Cited by 29
González, N., Hanalıoğlu, Ş., Mignucci-Jiménez, G., Abramov, I. T., Xu, Y., & Preul, M. C. (2022) · Operative Neurosurgery · doi:10.1227/ons.0000000000000544 · Cited by 28
Aggregating Feature Point Cloud for Depth Completion
Zhu, Y., Sheng, Z., Zhou, Z., Luo, L., Cao, S.-Y., Gu, H., Zhang, H., & Shen, H. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00802 · Cited by 29
“In recent studies, ViTs [8, 28] have demonstrated huge potential in various vision tasks due to larger receptive field, including image classification [3, 24], image segmentation [40, 25], dense prediction [33, 50], etc .”
BackgroundMethodDiffCAD: Weakly-Supervised Probabilistic CAD Model Retrieval and Alignment from an RGB Image
Gao, D., Rozenberszki, D., Leutenegger, S., & Dai, A. (2024) · ACM Transactions on Graphics · doi:10.1145/3658236 · Cited by 38
“While 2D semantic understanding from a single RGB image has achieved significant success in recent years [32, 40, 65, 68], 3D perception – required for enabling tasks such as navigation or interaction with objects – remains challenging, requiring not only object recognition but reconstruction of…”
BackgroundLi, G., Chi, X., & Qu, X. (2023) · Automotive Innovation · doi:10.1007/s42154-023-00223-6 · Cited by 32
“The supervised methods include those found in Bhat et al. [19], Song et al. [20], Ranftl et al. [21], Eigen et al. [46], Liu et al. [52] and Kundu et al. [53].”
MethodAutomatic Medical Image Segmentation with Vision Transformer
Zhang, J., Li, F., Zhang, X., et al. (2024) · Applied Sciences · doi:10.3390/app14072741 · Cited by 28
Zhou, W., Jian, B., Fang, M., Dong, X., Liu, Y., & Jiang, Q. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3382354 · Cited by 29
“Similar to [5] and [34], we used a normal data enhancement strategy with random scaling in the range of [0.9,1.1], random cropping, and random horizontal flipping such that all image sizes were processed to 480 × 640 pixels.”
BackgroundMethodInfluentialDELTAR: Depth Estimation from a Light-Weight ToF Sensor and RGB Image
Li, Y., Liu, X., Dong, W., Zhou, H., Bao, H., Zhang, G., Zhang, Y., & Cui, Z. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19769-7_36 · Cited by 29
RealFlow: EM-Based Realistic Optical Flow Dataset Generation from Videos
Han, Y., Luo, K., Luo, A., Liu, J., Fan, H., Luo, G., & Liu, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19800-7_17 · Cited by 32
A Transformer-Based Decoder for Semantic Segmentation with Multi-level Context Mining
Shi, B., Jiang, D., Zhang, X., Li, H., Dai, W., Zou, J., Xiong, H., & Tian, Q. C. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19815-1_36 · Cited by 30
DACFusion: Dual Asymmetric Cross-Attention guided feature fusion for multispectral object detection
Qian, J., Qiao, B., Zhang, Y., Liu, T., Wang, S., Wu, G., & Han, D. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.129913 · Cited by 27
NTIRE 2025 Challenge on HR Depth From Images of Specular and Transparent Surfaces
Ramirez, P. Z., Tosi, F., Stefano, L. D., Timofte, R., Costanzino, A., Poggi, M., Salti, S., Mattoccia, S., Zhang, Z., Yang, Y., et al. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00098 · Cited by 29
“It builds over the DPT backbone [79], enhanced through a metric bins module implemented for learning metric depth rather than an affine-invariant output.”
Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail
Bartolomei, L., Tosi, F., Poggi, M., & Mattoccia, S. (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/cvpr52734.2025.00103 · Cited by 58
DualFormer: Local-Global Stratified Transformer for Efficient Video Recognition
Liang, Y., Zhou, P., Zimmermann, R., & Yan, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19830-4_33 · Cited by 31
“Recently, Transformers [52] have presented an alternative visual modeling paradigm beyond CNNs, demonstrating great potential in many image processing tasks [36, 43, 53,55,57,64].”
BackgroundSabedotti, M. E. D. S., O’Regan, A. C., & Nyhan, M. (2023) · Environmental Science & Technology · doi:10.1021/acs.est.3c05000 · Cited by 30
SAMFusion: Sensor-Adaptive Multimodal Fusion for 3D Object Detection in Adverse Weather
Palladin, E., Dietze, R., Narayanan, P., Bijelic, M., & Heide, F. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-73030-6_27 · Cited by 38
“Those checkpoints were either based on the NuScenes [4] or Kitti [11] datasets.”
MethodIEBins: Iterative Elastic Bins for Monocular Depth Estimation and Completion
Shao, S., Pei, Z., Chen, W., Chen, P. C. Y., & Li, Z. G. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02293-3 · Cited by 26
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
