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 1,801–1,900 of 2,248
SeeClear: Reliable Transparent Object Depth Estimation via Generative Opacification
Wang, X., He, Y., Shi, J., Lu, J., Yang, Y., Jiang, Y., & Jiang, C. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37211-6_5
Najafli, E., Kästingschäfer, M., Bernhard, S., Brox, T., & Geiger, A. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-12840-9_35
Raj, S., Madhabhavi, B. A., Kumar, M., et al. (2026) · ACM Transactions on Cyber-Physical Systems · doi:10.1145/3828721
“Some methods [19] exhibit higher errors at longer ranges, or under poor/harsh lighting [22].”
SADAMB: Advancing Spatially-Aware Vision-Language Modeling Through Datasets, Metrics, and Benchmarks
Papadopoulos, G. T., Drakoulis, P., Ntovas, A., et al. (2025) · Computers · doi:10.3390/computers14100413
“DPT (Dense Prediction Transformer): Ranftl et al. [27] is another variant of the Vision Transformer, optimized for dense prediction tasks such as depth estimation and segmentation.”
Multiscale Attention-Enhanced Kolmogorov-Arnold Network for Concrete Porosity Segmentation
Hosseinzadeh, A., & Dehestani, M. (2026) · Journal of Computing in Civil Engineering · doi:10.1061/jccee5.cpeng-7709
Smart vehicule trajectory prediction in various autonomous driving scenarios
Gilles, T. (2023) · theses.fr (ABES) · doi:10.70675/2a784380z6b08z4daez964cz706f7921af08
Monocular Depth Estimation: Model Adaptation Evaluation on Synthetic-to-Real Transfer
Duc, M. N., Le-Dien, B., Le-Dang, H., et al. (2026) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-3-032-23407-0_2
UniLiPs: Unified LiDAR Pseudo-Labeling with Geometry-Grounded Dynamic Scene Decomposition
Ghilotti, F., Brucker, S., Saidy, N., Matteucci, M., Bijelic, M., & Heide, F. (2026) · International Conference on 3D Vision (3DV) · doi:10.1109/3dv69130.2026.00160 · Cited by 1
“Conversely image-based depth foundation models [4–6, 34, 45, 46, 48, 49, 63, 64] have demonstrated significant potential for generating dense depth predictions from single images but still lack behind when delivering metric depth accuracy.”
Transferable Attacks for Semantic Segmentation
He, M., Zhang, J., & Yu, X. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-1242-0_28
Salient Object Segmentation in 360° images/videos and light field
Zhang, Y. (2022) · theses.fr (ABES) · doi:10.70675/7b4980a9z109dz4ccez8cf5z0fef8b2c9cb6
Multimodal Monocular Geometric Clues-Informed Multi-View Stereo Network
SuI, W., & Xiong, S. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00679
“The adoption of Vision Transformer architectures, including DPT [27] and Omnidata [8], has further improved the performance of the model.”
Depth Estimation Through Sub-Pixel Laplacian Pyramid NetworkPreprint
Ali, A., Adibi, P., & Ehsani, S. M. S. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4855393
Na, H., Liu, D., & Wang, S. (2023) · Electronics · doi:10.3390/electronics12030763
A Mixture-of-Experts Framework Based on Depth Images for Text to Video Storyboard Task
Gu, X., Ni, F., & Song, R. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-5740-0_33
A Unified Framework for Depth-Assisted Monocular Object Pose Estimation
Hoang, D.-C., Tan, P. X., Nguyen, T.-U., Pham, H.-N., Nguyen, C.-M., Bui, S.-A., Duong, Q.-T., Vu, V.-D., Nguyen, V.-T., Duong, V.-H., et al. (2024) · IEEE Access · doi:10.1109/access.2024.3443148
“ViT-based ViT-based methods include Adabins [43], DPT [74], Shim et al. [75], YOLOPose [76], Trans6d [77], and Zhou et al [78].”
MethodDuetGS: Two-Stage Controllable 3D Human Reconstruction From Dual Images
Chen, Y., Wan, B., Shi, Y., et al. (2026) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2026.3685443
“Sparse view reconstruction: In some earlier NeRF-based sparse reconstruction works [38], [39], [40], [41], [42], [43], [44], [45], visibility or depth information has been utilized to guide the reconstruction process.”
Tiny-Dataset Scale-Consistent Diffusion Framework for Single SAR Image Building Height Estimation
Li, Q., Wu, Y., Zhang, H., et al. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3693803
“Model capacity has steadily increased, from relatively modest CNNs to large transformer-based architectures [18], [19].”
MonoSOWA++: Universal Autolabeling Pipeline for Monocular 3D Object Detection
Skvrna, J., & Neumann, L. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-03007-7
Accurate Polyp Sizing via Attention-Guided Parallel CNN-ViT Depth and Learned Scale
Nabijiang, A., Liu, F., Feng, J., Chen, Q., Cao, Y., & Liu, B. (2025) · International Conference on Machine Learning and Applications · doi:10.1109/icmla66185.2025.00031
AerialNet: Monocular Depth Estimation Architecture for UAV Aerial Images Using ViT & GraphSAGE
Narayan, I., Dhaliwal, R. P. S., Singh, S., & Poddar, S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-18477-1_26
Semantic Communications With World Models
Jiang, P., Guo, J., Wen, C.-K., et al. (2026) · IEEE Transactions on Communications · doi:10.1109/tcomm.2026.3714287
“The average performance degradation curves obtained from these scenarios are used as the corresponding L References.”
On the relevance of patch-based extraction methods for monocular depth estimation
Coscia, P., Fusillo, A., Genovese, A., Piuri, V., & Scotti, F. (2025) · Image and Vision Computing · doi:10.1016/j.imavis.2025.105857
Visionary Co-Driver: Enhancing Driver Perception of Potential Risks With LLM and HUD
Xiang, W., Lei, Z., Wang, J., Huang, Y., Zheng, Q., Zhang, T., Zhao, A., & Sun, L. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3634580
“Given that the DPT model, utilized for depth estimation, has demonstrated strong performance on the ADE20K and Pascal Context datasets, even with smaller datasets yielding excellent results, we did not conduct additional evaluations [36].”
Self-Supervised Monocular Depth Estimation with Multi-Scale Wavelet Enhancement
Wang, S. W., Wu, J., Lu, Q., et al. (2025) · IEEE International Conference on Robotics and Biomimetics (ROBIO) · doi:10.1109/robio66223.2025.11378264
“Several works have introduced Transformers [12] to capture dependencies between distant regions.”
Improved Classification-Based Monocular Depth Estimation for Robust Generalization of Diverse Scenes
Yang, W., Wu, G.-W., & Yang, J. (2026) · Applied Sciences · doi:10.3390/app16168003
Combining Projected Uncertainty for Self-Supervised Visual Odometry: From Two-Frame to Multi-Frame
Xie, J., Dhaouadi, O., Chen, W., Meier, J., Bauer, Z., Pollefeys, M., & Cremers, D. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02915-y
Compacting Singleshot Multi-Plane Image via Scale Adjustment
Bergfelt, M., Larsson, V., Saitô, H., & Mori, S. (2023) · IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct) · doi:10.1109/ismar-adjunct60411.2023.00117
UAV 3D Scene Understanding: A Survey from an Agent-Capability Evolution Perspective
Zhu, E., Xu, L., Chen, Z., Cui, J., Wang, J., Zou, Y., Yang, K., Liu, X., Qi, X., & Wang, L. (2026) · Remote Sensing · doi:10.3390/rs18142323
Development of a Vision System for Determining Geometric Parameters of Objects
Arkhipov, A., & Kravchuk, V. V. (2025) · International Ural Conference on Electrical Power Engineering (UralCon) · doi:10.1109/uralcon67204.2025.11206674
“Depth Anything V2 leverages a hierarchical vision transformer (ViT) [12] backbone with multi-scale feature fusion, implementing a Dense Prediction Transformer (DPT) [13] decoder to generate high-resolution depth maps.”
M33D: Learning 3D priors using Multi-Modal Masked Autoencoders for 2D image and video understanding
Jamal, M. A., & Mohareri, O. (2024) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv57701.2024.00253
“Following MultiMAE, we use DPT [55] as dense prediction head on the top of the ViT encoder.”
MethodA Novel Method for Monocular Depth Estimation Using an Hourglass Neck Module
Oh, S.-J., & Lee, S.-H. (2024) · Sensors · doi:10.3390/s24041312
Sato, S., Murasaki, K., & Tanida, R. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-31583-0_40
Real-Time Lightweight Monocular Depth Estimation Using Dilated-Strip Feature Enhancement
Hu, D., Zhang, H., Gu, J., et al. (2026) · doi:10.1109/caos69069.2026.11666327
FGS-SLAM: A Calibration-Free Slam Framework Using Feed-Forward 3D Gaussian Splatting
Yang, X., Cao, T., & Liu, Y. (2026) · IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW) · doi:10.1109/vrw70859.2026.00298
“Given an image pair, our model encodes the input with a Vision Transformer (ViT) and applies DPT-style heads [4] on the decoder features for dense 3D reconstruction.”
Gao, S., Zhou, F., & Shao, S. (2026) · Electronics · doi:10.3390/electronics15174046
A Diffusion-Based Neuron Coverage Feedback Fuzz Testing Method
Yang, K., Wu, J., Cui, Y. J., & Li, G. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-9911-7_24
LET: Vision Transformer based Refinement Network for Light Field Editing
Jo, S.-U., Kim, G.-J., & Rhee, C. E. (2022) · International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC) · doi:10.1109/itc-cscc55581.2022.9894877
“Second, the visual quality is consistently improved through the refinement network which is based on the dense prediction transformer (DPT).”
BackgroundMethodInfluentialHoloQRam: Efficient Real-Time Spatial Video Delivery via Animated QR Codes
Choi, J., & Ko, J. (2026) · IEEE International Conference on Pervasive Computing and Communications · doi:10.1109/percom67906.2026.11524555
Lei, Y., Cao, L., Zhao, Y., et al. (2026) · Buildings · doi:10.3390/buildings16112212
A survey: advances in multi-modal visual understanding and generation
Xi, D., Zhou, Z., Wang, Y., Wang, J., & Zhang, C. (2026) · Vicinagearth. · doi:10.1007/s44336-025-00032-x · Cited by 1
“Generative Understanding Depth Prediction DPT [106], Marigold [2], DepthFM [107], DepthCrafter [108] Discriminative models; Refor-mulated as latent diffusion tasks Strong zero-shot performance; Temporally consistent sequences Surface Normals StableNormal [3], Normal-Crafter [109] Coarse-to-fine…”
Indoor Scene Reconstruction From Monocular Video Combining Contextual and Geometric Priors
Wen, M., Sheng, X., & Cho, K. (2024) · IEEE Access · doi:10.1109/access.2024.3481250
“Originally developed for natural language processing, the transformer model [14] has gained significant attention and achieved remarkable success across various computer vision tasks [15], [16] in recent years.”
BackgroundMethodAdaptive Valve Selection Algorithm for PET Bottle Sorting Equipment Based on Segment Anything Model
Li, C., Liu, L., Zhao, Y., et al. (2025) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc65474.2025.11091266
Image Segmentation Applied to Urban Surface and Aerial Constraints Analysis
Trani, M. L. A., & Madaschi, F. (2023) · Proceedings e report · doi:10.36253/10.36253/979-12-215-0289-3.90
Chen, Y., Pan, X. Y., Zhang, H., et al. (2026) · Journal of Building Engineering · doi:10.1016/j.jobe.2026.116066
Vision Based Predictive Collision Avoidance System for Navigation of Micro Aerial Vehicle
P., D., S, K., Anbarasu, B., et al. (2026) · 2026 4th International Conference on Artificial Intelligence and Machine Learning Applications Theme: Healthcare and Internet of Things (AIMLA) · doi:10.1109/aimla67915.2026.11522568
Graph-Based Similarity of Deep Neural NetworksPreprint
Chen, Z., Lu, Y., Hu, J., et al. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4758068
SoftCut: A Fully Differentiable Relaxed Graph Cut Approach for Deep Learning Image Segmentation
Bonfiglio, A., Cannici, M., & Matteucci, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-53969-5_37
Improving Zero-Shot Template-Based 6D Pose Estimation with Geometric Features
Pöllabauer, T., Weyel, J., Knauthe, V., Berkei, S., & Kuijper, A. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-77392-1_4
Qu, Y., Hu, X., Chen, F., & Wei, Z. (2026) · Applied Intelligence · doi:10.1007/s10489-026-07143-y
“Figure 20 illustrates typical failure cases under challenging illumination for DPT [ 52] and DepthAnything [53], including: Case 1, screen-plane mis-identification; Case 2, glass-surface reflection interference; Case 3, instability on small distant objects; Case 4, foreground and background…”
Evaluating Ground Imagery for Long-Standoff, Vision-Based Navigation, Localization and Positioning
Ruby, J. G., Carter, J. R., Pham, M. V., Shuart, W., Massaro, R. D., Fischer, R. L., & Anderson, J. E. (2026) · Applied Sciences · doi:10.3390/app16157397
Zhang, S., Wang, X., Li, J., et al. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755163
“Vision Transformers (ViTs) have recently emerged as promising alternatives, thanks to their global self-attention mechanisms that capture long-range dependencies effectively [14, 24, 36].”
Topological Regularization for Dense Prediction
Fu, D., & Nelson, B. J. (2022) · International Conference on Machine Learning and Applications · doi:10.1109/icmla55696.2022.00014
“Outside of the family of convolutional neural networks, Ranftl et al. [42] extends the vision transformers [16] to monocular depth estimation, and achieves better results than CNNs.”
BackgroundVisual Place Recognition Based on an Adaptive D-Value Optimization Strategy
Jian, Y.-H., & Lee, J. (2026) · Sensors · doi:10.3390/s26092799
Tăng cường huấn luyện mô hình học tự chú ý cho phân tích và phân đoạn tiếng nói
Tân, H. M., & Quốc, N. K. (2024) · Journal of Science and Technology · doi:10.55401/8vq83773
Kwon, B., & Kim, M. (2024) · Journal of Broadcast Engineering · doi:10.5909/jbe.2024.29.6.866
Spatially-variant Blur Degradation Model Based on Depth Estimation
Xie, R. J., Guo, S., Zou, L., et al. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888678
“Subsequently, it uses a pretrained DPT [18] model for depth estimation.”
Liu, Y., Jin, G., Zhang, J., et al. (2025) · International Conference on Generative Artificial Intelligence and Information Security (GAIIS) · doi:10.1145/3728725.3728800
“Focusing on urban greening research, cutting-edge technologies have formed a complete technology chain: from data acquisition, feature extraction (DeepLabV3+, DPT, and other models) [3, 4], to spatial analysis (Floyd-Warshall algorithm path optimization) [5].”
Learning Motion Trends in Gaussian Splatting for Monocular Dynamic Reconstruction
Xie, R., Hu, B., Cui, Y., et al. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463323
Deep Learning-Based Monocular Depth Estimation in Cultural Heritage
Arampatzakis, V., Arnaoutoglou, F., Koutsoudis, A., et al. (2025) · Springer eBooks · doi:10.1007/978-3-032-06389-2_15
Fan, C.-L., Cao, M., Hung, C. C., et al. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00606
“We evaluate several deep learning architectures for sharpness regression, including VIT [27], U-Net [28], Res-Unet [36], Dense-Unet, and U2-Net [26].”
Comparing YOLO Models for Real-Time ASL Recognition
Gomez, A., & Arzuaga, E. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-032-09694-4_17
Multi-path Segmentation Network Based on CNN and Transformer for Skin Lesion Image
Nie, T., Zhao, Y., & Yao, S. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0963-5_23
Advancements in Machine Learning-Based Depth Estimation
He, H. (2024) · Highlights in Science Engineering and Technology · doi:10.54097/f8y4ta42
S-VAM: Shortcut Video-Action Model by Self-distilling Geometric and Semantic Foresight
Yan, H., Zhong, Z., Zhu, J., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37035-8_36
Stereo3D-NeRF: Generating 3D Visualizations with Paired Stereoscopic Views
Wang, Y., Zhou, G., Liu, W., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-6950-2_16
Maxwell, A. E., Ramezan, C. A., & He, Y. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-29306-1.00016-8
Discrete Diffusion Propagated Transformer For Flexible Ureteroscopic Semantic Segmentation
Du, X., Yang, M., Luo, G., et al. (2025) · Proceedings - International Conference on Image Processing · doi:10.1109/icip55913.2025.11084557
An Efficient Hybrid Network for Tibial Dyschondroplasia Segmentation: Application to Medical Images
Yang, Y., Liu, X., Feng, C., et al. (2025) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm66473.2025.11356847
DS$$^{2}$$Net: Joint Depth and Segmentation Learning with Synthetic Supervision
Cao, C.-T., Le, T.-D., Mai-Ton, D.-K., & Le, T.-S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-36871-3_14
A Hybrid ResNet U-Net++ Architecture with ASPP and SE for Fish Histological Image Segmentation
Silva, A. F. M., Cruz, Y. L. K. F., Braga, K. R., Araujo, W. B. D. D., Neta, R. N. F. C., & Santana, E. É. C. (2026) · Eng—Advances in Engineering · doi:10.3390/eng7070310 · Cited by 2
Huang, Y.-J., & Li, K. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5405061
Hybrid Dual Volume Learning for Iterative Fusion and Adaptive Depth Refinement for Shape from Focus
Ashfaq, K., & Mahmood, M. T. (2026) · Mathematics · doi:10.3390/math14101619
Landscape Semantic Segmentation: A Method Based on Transformer with Residual Refinement
Luan, K., & Yang, M. (2026) · International Conference on Neural Networks, Information and Communication Engineering (NNICE) · doi:10.1109/nnice68970.2026.11466236
“Vision Transform-ers address this by capturing long-range interactions, as seen in ViT and Swin Transformer [4], [5].”
“Separate-coupled” hierarchical framework for accurate visual object tracking
Liu, L., Liu, K., Wei, Z., et al. (2026) · Image and Vision Computing · doi:10.1016/j.imavis.2026.105982
Semantic-Guided 3D Gaussian Splatting for Sparse View Reconstruction and Segmentation
Padnekar, M., Mitra, K., & Das, S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-31927-2_14
Window Token Concatenation for Efficient Visual Large Language Models
Li, Y., Bao, W., Ye, B., Tan, Z., Chen, T., Liu, H., & Kong, Y. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00301 · Cited by 1
“The Vision Transformer (ViT) [52] is extensively utilized in numerous vision tasks [16, 27, 43, 50, 53], but it has long struggled with quadratic complexity.”
A Multi-ground Truth Approach for RGB-D Saliency Detection
Huynh, N. V., Pham, V. L., Mai, X. T., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78447-7_2
Xiang, X., Yan, Q., Qiao, Y., & Saddik, A. E. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4768854
Wu, Z., Ou, W., Zheng, Q., Wu, Q., Gao, D., Fang, W., Chen, C., Yang, Y., & Wang, Z. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110751
Late sensor fusion approach with a designed multi-segmentation network
Çaldıran, B. E., & Acarman, T. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-10004-9
“Dense prediction transformer (DPT) in [27] utilizes vision transformers instead of convolutional networks as a backbone for dense prediction tasks.”
Monocular Depth Estimation Using Differential Scale Features and Log-Remapping Depth Guidance
Ou, X., & Wu, Y. (2024) · International Conference on Computing and Pattern Recognition (ICCPR) · doi:10.1145/3704323.3705010
“With the success of ViT on other tasks, Yang et al.[30] and Ranftl et al.[25] used transformer instead of CNN to obtain better monocular depth estimation performance.”
AI-Based QRS Onset Detection in the Early Ventricular Activation Site ECGs
Reznichenko, S., Whitaker, J., Ni, Z., et al. (2025) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2025.3605298
“Transformer architecture has surpassed CNNs in various areas [33], [34], [35], which could potentially improve the accuracy of QRS onset detection.”
Hierarchical Feature Concatenation Fusion Network for Tooth Image SegmentationPreprint
Li, B., Kong, L., & Zhao, M. (2024) · Research Square · doi:10.21203/rs.3.rs-5354031/v1
StereoGlue: Robust Estimation with Single-Point Solvers
Baráth, D., Mishkin, D., Cavalli, L., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72998-0_24
An Improved Diffusion-Based Method with Depth Guidance for Monocular Depth Estimation
Yan, Z., Yan, B., Chen, Y., & Xu, W. (2025) · IEEE International Conference on Advanced Infocomm Technology · doi:10.1109/icait66450.2025.11353306
“This trend has evolved from convolutional encoder-decoder architectures toward increasingly powerful vision transformers [6], benefiting from larger model capacities and larger-scale data.”
Guided Querying over Videos using Autocompletion Suggestions
Yoo, H., & Nandi, A. (2024) · Workshop on Human-In-the-Loop Data Analytics (HILDA) · doi:10.1145/3665939.3665964
“There is a significant body of work focused on analyzing videos for tasks such as detection [19, 42, 43], tracking [4, 47], and estimating 3D locations [14, 41].”
BackgroundWRD-Seg:Water Reflection-suppressed depth-guided semantic segmentation for surface scenes
Zhou, M., Liang, Z., & Dong, W. (2026) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2026.200708
CF 3 : Compact and Fast 3D Feature Fields
Lee, H., Min, J., & Park, J. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02590
“LSeg, based on the DPT [37] backbone, and CLIP with SAM both produce features at nearly the same resolution as the input image.”
Chen, J. (2024) · International Conference on Cyber Security and Information Engineering (ICCSIE) · doi:10.1145/3689236.3696753
“Lightweight models significantly reduce the computational resource consumption of the model by minimizing redundant computation paths and compressing the network structure [8-10].”
Mou, C., Fan, J., Liu, A., et al. (2026) · Agriculture · doi:10.3390/agriculture16151656
Tucker, R., Chao, A. F., Rim, N., Gaillard, E. P., Akcelik, G. N., & Berman, M. G. (2025) · Journal of Quantitative Criminology · doi:10.1007/s10940-026-09675-5
Recovering Dense Metric Depth in Indoor Scenes from Monocular Depth Foundation Models and 2D LiDARs
Marsal, R., Chapoutot, A., Xu, P., & Filliat, D. (2025) · Springer proceedings in advanced robotics · doi:10.1007/978-3-031-89471-8_36
TransIndoor: Transformer Based Self-supervised Indoor Depth Estimation
Zhang, H., Li, Z., Geng, Y., et al. (2024) · Lecture notes in electrical engineering · doi:10.1007/978-981-97-2757-5_61
CheapNVS: Real-Time On-Device Narrow-Baseline Novel View Synthesis
Georgiadis, K., Yucel, M. K., & Saà-Garriga, A. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10888972
“We use DPT [28] depth maps for COCO training to be fair against AdaMPI, and Marigold [29] depth maps for Open Images training.”
Lei, 翟. Z., Xiang, 黄. H., Kai, 孟. M., et al. (2026) · Chinese Journal of Lasers · doi:10.3788/cjl260657
Dense Monocular Slam Using Photometric Calibration and ZoedepthPreprint
Song, D., Zuo, T., & Mei, J. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4846951
FMDP: Leveraging a Foundation Model for Dual-Pixel Disparity Estimation
Lee, D., Wu, Z., Monno, Y., & Okutomi, M. (2025) · doi:10.23919/mva65244.2025.11175082
“This assembly process is motivated by Dense Prediction Transformer’s Reassemble module [14].”
Zhang, W., Li, J., Jin, X., et al. (2026) · Youth Academic Annual Conference of Chinese Association of Automation (YAC) · doi:10.1109/yac71005.2026.11615971
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
