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 901–1,000 of 2,248
Object Pose Estimation Using Color Images and Predicted Depth Maps
Hoang, D.-C., Tan, P. X., Nguyen, A.-N., Vu, D.-Q., Vu, V.-D., Nguyen, T.-U., Duong, Q.-T., Nguyen, V.-T., Hoang, N.-A., Phan, K.-T., et al. (2024) · IEEE Access · doi:10.1109/access.2024.3397715 · Cited by 5
“The monocular depth estimation configurations align with those detailed in the original papers [73], [74].”
BackgroundMethodSuperPrimitive: Scene Reconstruction at a Primitive Level
Mazur, K., Bae, G., & Davison, A. 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/cvpr52733.2024.00476 · Cited by 5
“Depth prediction [12, 35, 36] is the most obvious choice but leads to a rigid per-image reconstruction which is difficult to feed into multi-view optimisation.”
BackgroundMethodPerceived distance to greenery affects psychological restoration
Ma, H., Zhang, Y., Cui, Q., Kwan, M., & Cai, J. (2025) · Environment and Planning B Urban Analytics and City Science · doi:10.1177/23998083251362610 · Cited by 5
Recognition of Apple Leaves Infection Using DenseNet121 with Additional Layers
Nain, S., Mittal, N., & Jain, A. (2024) · Lecture notes in networks and systems · doi:10.1007/978-981-99-9562-2_24 · Cited by 5
A Physical Model-Guided Framework for Underwater Image Enhancement and Depth Estimation
Du, D., Si, L., Xu, F., Niu, J., & Sun, F. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3674091 · Cited by 9
“Depth Anything employs a DINOv2 encoder [53] for feature extraction and a dense prediction Transformer (DPT) decoder [54] to predict disparity, i.e., relative inverse depth.”
Erişen, S., Mehranfar, M., & Borrmann, A. (2025) · Remote Sensing · doi:10.3390/rs17162910 · Cited by 5
Learning Cross-modality Interaction for Robust Depth Perception of Autonomous Driving
Liang, Y., Chen, N., Yu, Z., Tang, L., Yu, H., Guo, B., & Zeng, D. D. (2024) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3650039 · Cited by 5
“For the encoder, variants of CNN (e.g., ResNet [27], DeepLab [7], and Unet [59]) and Transformer (e.g., ViT [20, 56, 82]) are used to learn the high-quality feature maps including the geometrical structures and boundary information from in the imagery [82].”
MethodXing, Z., Liu, L., Yang, Y., Wang, H., Tian, Y., Chen, S., Li, W., Liu, G., & Zhu, L. (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/cvpr52734.2025.02372 · Cited by 5
KUŞ, İ., Koçak, C., & Keleş, A. (2025) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2025.200615 · Cited by 11
Reinforcing Local Structure Perception for Monocular Depth Estimation
Zhang, S., Dong, J., Chervan, A., Kurlovich, D. M., Hou, W. W., & Ding, M. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3293156 · Cited by 5
“Meanwhile, considering the impressive depth estimation approaches based on transformer [40], [41], [42], we also utilize SwinTransformer-Large (SwinL) [43] as the backbone when comparing to SOTA methods.”
MethodTransLevelSet: Integrating vision transformers with level-sets for medical image segmentation
Koutsiou, D.-C. C., Koutsiou, D.-C. C., Savelonas, M. A., et al. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128077 · Cited by 5
Bridging local and global representations for self-supervised monocular depth estimation
Lin, M., Li, G., & Hao, Y. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.108277 · Cited by 6
Structure-Centric Robust Monocular Depth Estimation via Knowledge Distillation
Chen, R., Luo, H., Zhao, F., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0969-7_8 · Cited by 5
Mucha, W., Wray, M., & Kampel, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78354-8_12 · Cited by 5
FlowSeek: Optical Flow Made Easier with Depth Foundation Models and Motion Bases
Poggi, M., & Tosi, F. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00537 · Cited by 6
“In depth estimation, models like Depth Anything [92] and its successor Depth Anything v2 [93] have achieved state-of-the-art results through synthetic data training and knowledge distillation, alongside other VFMs for monocular depth estimation [4, 39, 58, 59].”
Egocentric Scene Understanding via Multimodal Spatial Rectifier
Do, T., Vuong, K., & Park, H. S. (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/cvpr52688.2022.00285 · Cited by 6
“(9) MiDaS [45], (10) DPT [44]: state-of-the-art depth prediction model trained on a large scale dataset MIX6 [44].”
BackgroundWu, X., Wei, X., Xu, H., He, W., Zhou, L., Hou, Y., & Wang, Y. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3427794 · Cited by 7
“In this article, DPT [44], a transformer framework for computer vision, is used as backbone, and CNN is used as an adapter to introduce image-related inductive biases into backbone.”
MethodDepthGAN: GAN-based depth generation from semantic layouts
Li, Y., Xiao, J., Wang, Y., & Lu, Z. (2024) · Computational Visual Media · doi:10.1007/s41095-023-0350-8 · Cited by 10
“We tried to retrain depth estimation models, such as Adabins [3], Midas [42] and DPT [41] using a semantic layout as input, but do not receive satisfying output.”
Ceccato, V. A., Kang, Y., Abraham, J., Näsman, P., Duarte, F., Gao, S., Ljungqvist, L., Zhang, F., & Ratti, C. (2025) · The British Journal of Criminology · doi:10.1093/bjc/azaf017 · Cited by 5
Global-Local Self-Distillation for Visual Representation Learning
Lebailly, T., & Tuytelaars, T. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00149 · Cited by 7
“Monolithic vision transformers [17] have induced a wave of work on multi-stage vision trans-formers [31, 51, 44, 38] which do not process patch tokens at a single resolution ( e.g .”
BackgroundImage Appeal Revisited: Analysis, New Dataset, and Prediction Models
Göring, S., & Raake, A. (2023) · IEEE Access · doi:10.1109/access.2023.3292588 · Cited by 5
“the deep learning approach described in [74] and [75] are shown.”
Lech, P. (2025) · Electronics · doi:10.3390/electronics14163204 · Cited by 5
TRICKY 2024 Challenge on Monocular Depth from Images of Specular and Transparent Surfaces
Ramirez, P. Z., Costanzino, A., Tosi, F., Poggi, M., Stefano, L. D., Weibel, J.-B., Bauer, D., Antensteiner, D., Vincze, M., Li, J., et al. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-91569-7_16 · Cited by 5
RenderFormer: Transformer-based Neural Rendering of Triangle Meshes with Global Illumination
Zeng, C., Dong, Y., Peers, P., Wu, H., & Tong, X. (2025) · ACM SIGGRAPH Conference (SIGGRAPH) · doi:10.1145/3721238.3730595 · Cited by 15
“Furthermore, we employ a dense vision transformer [Ranftl et al. 2021] To reduce the degrees of freedom in the training data, we perform the view-dependent stage in camera coordinates, rather than in world coordinates as in the view-dependent stage.”
A fully locally selective large kernel network for traffic video detection
Hou, Y., Zhang, Z., Du, L., & Yin, J. (2024) · Measurement · doi:10.1016/j.measurement.2024.115779 · Cited by 5
Qu, F., Sun, Y., Zhou, M., Liu, L., Yang, H., Zhang, J., Huang, H., & Hong, D. (2023) · Remote Sensing · doi:10.3390/rs16010003 · Cited by 4
Towards Cross-View-Consistent Self-Supervised Surround Depth Estimation
Ding, L., Jiang, H., Li, J., Chen, Y., & Huang, R. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802436 · Cited by 6
“Recent years have witnessed remarkable development of image-based depth estimation [1], [2], [3], [4], [5] and its applications in various scenarios, including 3D object detection [6], [7] and BEV segmentation [8], [9], etc.”
Lifelong-MonoDepth: Lifelong Learning for Multidomain Monocular Metric Depth Estimation
Hu, J., Fan, C., Zhou, L., Gao, Q. H., Liu, H., & Lam, T. L. (2023) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2023.3323487 · Cited by 7
“Most previous approaches [29], [28], [36] of multi-domain learning choose only to infer relative depth maps to tackle domain gaps.”
BackgroundShareGS: Hole completion with sparse inputs based on reusing selected scene information
Hong, Y., Tao, B., & Gong, Z. (2025) · Pattern Recognition · doi:10.1016/j.patcog.2025.111729 · Cited by 5
PCformer: A parallel convolutional transformer network for 360° depth estimation
Xu, C., Yang, H., Han, C., & Zhang, C. (2022) · IET Computer Vision · doi:10.1049/cvi2.12144 · Cited by 5
Unsupervised optical flow estimation method based on transformer and occlusion compensation
Xiang, X., Abdein, R., & Lv, N. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07483-z · Cited by 5
“Unlike some computer vision tasks, such as image classi-fication, where feature resolution and granularity may not matter, the quality of the extracted features is critical and required for dense prediction tasks [1].”
BackgroundMethodMulti-Task Self-Training for Learning General RepresentationsPreprint
Ghiasi, G., Zoph, B., Cubuk, E. D., Le, Q. V., & Lin, T.-Y. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2108.11353 · Cited by 5
V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learning
Mur-Labadia, L., Muckley, M. J., Bar, A., Assran, M., Sinha, K., Rabbat, M., LeCun, Y., & Ballas, N. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37369-4_36 · Cited by 150
LoGS: Visual Localization via Gaussian Splatting with Fewer Training Images
Cheng, Y., Jiao, J., Wang, Y., et al. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11127759 · Cited by 5
“When ground truth depth is absent, we generate monocular depth maps ˆ D for training images using the pre-trained Dense Prediction Transformer (DPT) [35] to regularize the training.”
Multi-View Depth Estimation by Using Adaptive Point Graph to Fuse Single-View Depth Probabilities
Wang, K., Liu, C., Liu, Z., Xiao, F., An, Y., Zhao, X., & Shen, S. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3405332 · Cited by 5
“And meth-ods [1], [2] utilize the vision transformer to encode the global context features and regress depth from the global features.”
BackgroundMethodPseudo-Viewpoint Regularized 3D Gaussian Splatting For Remote Sensing Few-Shot Novel View Synthesis
Wang, Y., Tang, X., Ma, J., et al. (2024) · IEEE International Geoscience and Remote Sensing Symposium (IGARSS) · doi:10.1109/igarss53475.2024.10640765 · Cited by 5
“One straight-forward insight is to introduce the depth estimation network [5] to estimate object depth [6], which can provide direct geometric information [7].”
BackgroundSemantic discrete decoder based on adaptive pixel clustering for monocular depth estimation
Liu, X., Tang, S., Ye, M., et al. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107565 · Cited by 5
Reconfigurable Intelligent Surface Aided Wireless Sensing for Scene Depth Estimation
Taha, A., Luo, H., & Alkhateeb, A. (2023) · International Conference on Communications · doi:10.1109/icc45041.2023.10279208 · Cited by 5
“(a) The scene under study; (b-c) the RGB based estimated maps [10], [11]; (d) the ground truth map; (e-f) the RIS-based estimated maps.”
BackgroundMethodInfluentialA Layered Framework for Universal Extraction and Recognition of Electrical Diagrams
Cao, W., Chen, Z., Wu, C., et al. (2025) · Electronics · doi:10.3390/electronics14050833 · Cited by 5
RS-AGENT: Large Language Models Guided Agent System for Remote Sensing Image Generation
Zhu, L., Wu, J., Wang, B., Zhang, G., Wang, J., Chen, S., & Tan, M. (2024) · IEEE International Geoscience and Remote Sensing Symposium (IGARSS) · doi:10.1109/igarss53475.2024.10640804 · Cited by 5
“The methods employed for feature extraction encompass straight line [11], sketch [12], depth [13], normal map [14], edge [15], and soft edge [12].”
MethodAltNeRF: Learning Robust Neural Radiance Field via Alternating Depth-Pose Optimization
Wang, K., Yan, Z., Tian, H., Zhang, Z., Li, X., Li, J., & Yang, J. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i6.28360 · Cited by 5
Enhancing Diffusion Models with 3D Perspective Geometry Constraints
Upadhyay, R., Zhang, H., Ba, Y., Yang, E., Gella, B., Jiang, S., Wong, A., & Kadambi, A. (2023) · ACM Transactions on Graphics · doi:10.1145/3618389 · Cited by 9
“In order to evaluate the performance of the models, we follow the procedure used by [Ranftl et al. 2021] and we adopt common depth estimation metrics: Absolute relative error (Abs Rel), Square relative error (Sq Rel), Root mean squared error (RMSE), Log RMSE (RMSE log), and Threshold Accuracy ( 𝛿…”
BackgroundMethodInfluentialLightweight monocular depth estimation network for robotics using intercept block GhostNet
Ardiyanto, I., & Al-Fahsi, R. D. H. (2024) · Signal Image and Video Processing · doi:10.1007/s11760-024-03720-1 · Cited by 5
“In recent years, transformer network comes into the game too, e.g., [18], [19], and [20] used a vision-transformer for depth prediction.”
Composite Relationship Fields with Transformers for Scene Graph Generation
Adaimi, G., Mizrahi, D., & Alahi, A. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00014 · Cited by 4
“Such architectures have been applied to many visual tasks, such as classification [13, 48], detection [4, 14], semantic segmentation [58, 97], tracking [53], pose estimation [33] and human-object interaction [101, 28].”
MethodTADFormer: Task-Adaptive Dynamic TransFormer for Efficient Multi-Task Learning
Baek, S., Lee, S., Jo, H., Choi, H., & Min, D. (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/cvpr52734.2025.01384 · Cited by 6
“PEFT aims to adapt large-scale pre-trained models [5, 8– 10, 13, 19, 30, 30, 31, 31, 37–39, 46, 48, 53] to down-stream tasks by tuning only a small subset of parameters, without re-training the entire model.”
AutoColor: learned light power control for multi-color holograms
Zhan, Y., Kavaklı, K., Ürey, H., Sun, Q., & Akşit, K. (2024) · Optical Architectures for Displays and Sensing in Augmented, Virtual, and Mixed Reality (AR, VR, MR) · doi:10.1117/12.3000082 · Cited by 5
“To estimate the depth information for all the generated images at 2048 × 2048 resolution, we rely on a monocular depth information network 16,17 and their publicly available weights (dpt hybrid-midas-501f0c75.pt), consuming 35 GB memory on an NVIDIA A100 on a cloud GPU cluster.”
MethodSWAT: Spatial Structure Within and Among Tokens
Kahatapitiya, K., & Ryoo, M. S. (2023) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/106 · Cited by 10
“Token adoptation in vision tasks: At the moment, tokenbased models are widely applied in almost all domains in vision, including classification [22, 46, 65], object detection [6,16,90], segmentation [23,74], image generation [5,20,24, 38, 43], video understanding [1, 2, 4, 9, 25, 28, 41, 45, 47, 49, 56, 85], dense pre…”
Construction and Validation of a General Medical Image Dataset for Pretraining
Zhang, R., Pei, C., Shi, J., et al. (2024) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-024-01226-3 · Cited by 4
DenseSR: Image Shadow Removal as Dense Prediction
Lin, Y.-F., Lee, C.-M., & Hsu, C. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755664 · Cited by 4
Graph-Based Similarity of Deep Neural Networks
Chen, Z., Lu, Y., Hu, J., Xuan, Q., Wang, Z., & Yang, X. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128722 · Cited by 17
Dynamic Neural Radiance Field from Defocused Monocular Video
Luo, X., Sun, H., Peng, J., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72652-1_9 · Cited by 4
Morra, L., Santangelo, A., Basci, P., et al. (2024) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2024.104187 · Cited by 4
“2023) and its associated expert models (Ranftl et al., 2021; Poma et al., 2020; Zhou et al., 2022; Liu et al., 2018; Cheng et al., 2022a).”
BackgroundNadeem, N., Asad, M., Anwar, S., & Bais, A. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00539 · Cited by 6
“We obtain depth image estimations for all source and target images from Vision Transformer [39] via a pre-trained model trained on large-scale RGB data.”
InfluentialYang, X. Y., Lu, G., Zhang, J., Paneru, B., Dhungana, A., Dahal, S., Bist, R. B., & Chai, L. (2025) · Applied Sciences · doi:10.3390/app15126625 · Cited by 5
Multimodal Energy Prompting for Video Salient Object Detection
Jiang, T., Hou, F., & Wang, Y. (2024) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3696409.3700196 · Cited by 5
“For each video frame, we generate a synthetic depth map using DPT [50] within our MEPNet framework.”
POMATO: Marrying Pointmap Matching with Temporal Motions for Dynamic 3D Reconstruction
Zhang, S., Ge, Y., Tian, J., Xu, G., Chen, H., Chen, L., & Shen, C. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00538 · Cited by 5
“The specific number and the usage ratio of each dataset can be found in the supplementary materials.”
PCTDepth: Exploiting Parallel CNNs and Transformer via Dual Attention for Monocular Depth Estimation
Xia, C., Duan, X., Gao, X., et al. (2024) · Neural Processing Letters · doi:10.1007/s11063-024-11524-0 · Cited by 4
“Toaddressthesechallenges,deeplearning-basedneuralnetworks(suchasVGG,ResNet, PVT,andTransformer),thankstothegreatsuccessofusingConvolutionalNeuralNetworks (CNNs)[7],havebeenwidelyadoptedinMDEtasks[13,21,22]andhavemadesignificant progress.”
InfluentialSingle-Group Generalized RGB and RGB-D Co-Salient Object Detection
Wang, J., Yu, N., Zhang, Z., & Han, Y. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3514872 · Cited by 5
“…tasks, due to the lack of abundant paired data (RGB and depth images) to support training, existing methods resort to depth estimation algorithms [57] to generate corresponding depth data based on RGB CoSOD training sets (such as DUT-class [26], which comprises 291 groups and 8250 samples)…”
Artistic style decomposition for texture and shape editing
Reimann, M., Büßemeyer, M., Buchheim, B., et al. (2024) · The Visual Computer · doi:10.1007/s00371-024-03521-0 · Cited by 4
“InFig.7, we extract a depth map using DPT [49] which is used as a blending mask to blend between two parameter sets.”
Ma, Y., Chen, K., Tong, H., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72670-5_15 · Cited by 4
Sun, S., Tian, Y., Qi, Z., et al. (2022) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2022.104278 · Cited by 4
Calibrating Panoramic Depth Estimation for Practical Localization and Mapping
Kim, J., Lee, E. S., & Kim, Y. M. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00811 · Cited by 4
“[25], many existing works focus on developing neural network models that output depth maps from image input [26, 27, 28, 29, 30, 31].”
BackgroundMono‐MVS: textureless‐aware multi‐view stereo assisted by monocular prediction
Fu, Y., Zheng, M., Chen, P., & Liu, X. (2024) · The Photogrammetric Record · doi:10.1111/phor.12480 · Cited by 4
“Consistent with the mainstream dense prediction model, DPT (Ranftl et al., 2021) follows an encoder–decoder architecture.”
BackgroundResultInfluentialZhang, L. L., Xie, C., & Kitahara, I. (2025) · Journal of Imaging · doi:10.3390/jimaging11070246 · Cited by 4
“Subsequent work introduced monocular depth priors (e.g., MiDaS [11], DPT [22,23]) to compensate for missing depth information.”
Self-supervised monocular depth estimation with self-distillation and dense skip connection
Xiang, X., Li, W., Wang, Y., & Saddik, A. E. (2024) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2024.104048 · Cited by 4
SACNet: Saliency-Aided Aggregation Consensus Network for RGB-D Co-Salient Object Detection
Tu, Z., Qian, X., & Zhou, W. (2025) · IEEE Signal Processing Letters · doi:10.1109/lsp.2025.3567027 · Cited by 4
“To train the RGB-D Co-SOD method, we used a depth-estimation method [27] to generate corresponding depth maps.”
Sewer image super-resolution with depth priors and its lightweight network
Pan, G., Chen, W., Sui, Z., Guo, S., Lv, Y., Li, H., Sun, D., & Xia, Z. (2025) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.13453 · Cited by 5
“Building on the triumph of employing transformer in computer vision tasks, TransDepth (Zhao et al., 2021) and DPT (Ranftl, Bochkovskiy, & Koltun, 2021) have enhanced the performance of the model by replacing convolution operations with transformer network layers. and 𝐹 𝑠𝐿𝑅 into a…”
Chen, Y., Zhang, X., Chen, Y., et al. (2024) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme57554.2024.10687514 · Cited by 4
“For depth estimation, the pre-trained transformer backbone DINOv2 [18] is employed as the encoder, with the addition of reassemble and fusion stages [19] to enhance the fusion of features.”
MethodDynST: Dynamic Sparse Training for Resource-Constrained Spatio-Temporal Forecasting
Wu, H., Wen, H., Zhang, G., Xia, Y., Liang, Y., Zheng, Y., Wen, Q., & Wang, K. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3690624.3709391 · Cited by 13
MD2VO: Enhancing Monocular Visual Odometry through Minimum Depth Difference
Li, P., Tang, C., Duan, Y., & Li, Z. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10649955 · Cited by 4
Single-View Depth Estimation: Advancing 3D Scene Interpretation With One Lens
Dhanushkodi, K., Bala, A., & Chaplot, N. (2025) · IEEE Access · doi:10.1109/access.2025.3534324 · Cited by 4
“Ranftl et al. [1] demonstrated that Vision Trans-formers could outperform traditional convolutional networks in capturing long-range dependencies, which is crucial for accurate depth estimation.”
InfluentialLearning pose regression as reliable pixel-level matching for self-supervised depth estimation
He, J., Wang, Y., Wang, L., & Lu, H. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.130048 · Cited by 4
Segmentation of Non-Small Cell Lung Carcinomas: Introducing DRU-Net and Multi-Lens Distortion
Oskouei, S., Valla, M., Pedersen, A., Smistad, E., Dale, V. G., Høibø, M., Wahl, S. G. F., Haugum, M. D., Langø, T., Ramnefjell, M., et al. (2025) · Journal of Imaging · doi:10.3390/jimaging11050166 · Cited by 6
“The state-of-the-art in image segmentation tasks is the use of complex neural network architectures such as vision transformers and InternImage 16,17 .”
Revisiting classical DeepLab modernly for semantic segmentation
Lv, J., Huang, Y., Wan, X., & Duan, L. (2024) · International Conference on Image, Vision and Computing (ICIVC) · doi:10.1109/icivc61627.2024.10837090 · Cited by 4
ZiWen, D., Li, Y., & Ye, D. (2023) · Applied Intelligence · doi:10.1007/s10489-023-04872-2 · Cited by 4
“From Table 3, it can be observed that FasterMDE achieves slightly better accuracy compared to state-of-the-art vision-transformer-based methods, including SideRT [57], LapDepth [56], BTS [55], DPT [54], VNL [52], and TransDepth [53].”
MethodWu, B., Liu, C., Eckart, B., et al. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i3.20171 · Cited by 4
Encoder-Decoder based Segmentation Model for UAV Street Scene Images
Kumar, S., Kumar, A., Hong, H.-S., et al. (2023) · International Conference on Consumer Electronics · doi:10.1109/icce56470.2023.10043528 · Cited by 4
“After decent performance of transformers in Natural Language Processing (NLP) domain, it has gained huge attention in computer vision [1, 3, 4] area as well.”
BackgroundMaDis-Stereo: Enhanced Stereo Matching via Distilled Masked Image Modeling
Ahn, J., Choi, H., Kim, S., & Min, D. (2025) · IEEE Access · doi:10.1109/access.2025.3528022 · Cited by 5
“With the advancement of transformer frameworks [10], [11], [44], [45], [46], [47], stereo-matching models have evolved significantly, offering enhanced capabilities in stereo depth estimation tasks.”
SimGlim: Simplifying glimpse based active visual reconstruction
Jha, A., Seifi, S., & Tuytelaars, T. (2023) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv56688.2023.00035 · Cited by 6
“e constant) resolution has resulted in state-of-the-art results beating CNNs in dense prediction tasks such as semantic segmentation and depth prediction [22].”
BackgroundLightweight Self-Supervised Monocular Depth Estimation Through CNN and Transformer Integration
Wang, Z., Zou, Y., Lv, J., et al. (2024) · IEEE Access · doi:10.1109/access.2024.3494872 · Cited by 4
“Varma et al. [7] introduced a Dense Prediction Transformer [21] aimed at self-supervised monocular depth estimation and included an extra module to deduce the camera’s intrinsic parameters.”
MAPPNet: A Multi-Scale Attention Pyramid Pooling Network for Dental Calculus Segmentation
Nie, T., Yao, S., Wang, D., et al. (2024) · Applied Sciences · doi:10.3390/app14167273 · Cited by 4
Hybrid attention-guided RRT*: Learning spatial sampling priors for accelerated path planning
loulou, A., & Ünel, M. (2026) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2026.105338 · Cited by 4
Tozuka, R., Kadoya, N., Yasunaga, A., Saito, M., Komiyama, T., Nemoto, H., Ando, H., Onishi, H., & Jingu, K. (2025) · Japanese Journal of Radiology · doi:10.1007/s11604-025-01865-8 · Cited by 4
“DPT, originally using ViT-Hybrid or ViT-Large, was adapted using the more efficient Data-efficient image Trans-formers Tiny (DeiT-Tiny) backbone to avoid overfitting with our limited cases [22].”
Youwai, S., Chaiyaphat, A., & Chaipetch, P. (2024) · International Conference on Intelligent Computing and Next Generation Networks (ICNGN) · doi:10.1109/icngn63705.2024.10871286 · Cited by 4
Accurate ECG images classification using Vision Transformer
Alkegany, F. M. H., & Ali, K. H. (2024) · Basrah Researches Sciences · doi:10.56714/bjrs.50.1.26 · Cited by 4
SREGS: Sparse-view Gaussian radiance fields with geometric regularization and region exploration
Li, X., Li, K., Zhang, G., et al. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107820 · Cited by 4
Scale-Invariant Monocular Depth Estimation via SSI Depth
Miangoleh, S. M. H., Reddy, M. K. K., & Aksoy, Y. (2024) · ACM SIGGRAPH Conference (SIGGRAPH) · doi:10.1145/3641519.3657523 · Cited by 4
“Hence, modern MDE approaches are overwhelmingly data-driven to implicitly learn the depth cues [Eigen et al. 2014; Godard et al. 2017; Ramamonjisoa et al. 2020; Wang et al. 2020a; Wong and Soatto 2019; Zheng et al. 2018].”
Toward High-Resolution UAV Imagery Open-Vocabulary Semantic Segmentation
Chen, Z., Xie, Y., & Wei, Y. (2025) · Drones · doi:10.3390/drones9070470 · Cited by 4
The Fourth Monocular Depth Estimation Challenge
Obukhov, A., Poggi, M., Tosi, F., Arora, R. S., Spencer, J., Russell, C., Hadfield, S., Bowden, R., Wang, S., Ma, Z., et al. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00615 · Cited by 5
“To address this limitation, in the last three years, alongside to the development of more advanced frameworks and evaluation settings [81, 82], the Monocular Depth Estimation Challenge (MDEC) has pushed the research community toward more general models capable of cross-domain generalization,…”
InfluentialSIHENet: Semantic Interaction and Hierarchical Embedding Network for 360° Salient Object Detection
He, Z., Shao, F., Xie, Z., et al. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3507047 · Cited by 4
Impact of Pseudo Depth on Open World Object Segmentation with Minimal User Guidance
Schön, R., Ludwig, K., & Lienhart, R. W. (2023) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00509 · Cited by 4
“In order to be useful in downstream tasks, the MiDaS system ( [27,28]) has been trained on numerous datasets at the same time.”
On Moving Object Segmentation from Monocular Video with Transformers
Homeyer, C., & Schnörr, C. (2023) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw60793.2023.00095 · Cited by 4
“Since we only have monocular video as input, we create non-rgb pseudo-modalities with off-the-shelf expert networks for optical flow [54], depth [47] or scene flow [56].”
BackgroundMethodInfluentialTowards Practical Consistent Video Depth Estimation
Li, P., Ding, Y., Li, L., Guan, J., & Li, Z. (2023) · International Conference on Multimedia Retrieval · doi:10.1145/3591106.3592264 · Cited by 5
“For supervised approaches [26, 48, 49], since the training dataset and the data corresponding to the accurate depth map are difficult to obtain, these methods now use synthetic datasets, such as 3D movies [40], and consider a hybrid training strategy with multiple datasets to improve the generalization ability during…”
MethodSynthesizing Light Field Video from Monocular Video
Govindarajan, S., Shedligeri, P., Sarah, et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20071-7_10 · Cited by 4
“During inference, we first obtain the depth estimate z t from DPT[35] and convert it to a disparity map d t .”
BackgroundMethodInfluentialEfficient real-time semantic segmentation: accelerating accuracy with fast non-local attention
Lan, T., Dou, F., Feng, Z., & Zhang, C. (2023) · The Visual Computer · doi:10.1007/s00371-023-03135-y · Cited by 4
“DPT [37] leverages the transformer as an encoder, resulting in significant performance enhancements for dense prediction tasks.”
A robust and high-precision edge segmentation and refinement method for high-resolution images
Li, Q., & Chen, C. (2022) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2023049 · Cited by 4
Joint Edge and Regional Depth Enhancement Network for Camouflaged Object Detection
Miao, Q., Wang, Z., Liu, Y., et al. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889727 · Cited by 4
“To be specific, we utilize DPT [21] to generate depth maps for our training dataset.”
ConvNets vs. Transformers: Whose Visual Representations are More Transferable?Preprint
Zhou, H.-Y., Lu, C., Yang, S., et al. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2108.05305 · Cited by 4
“For depth estimation, we employ the scale-and shift-invariant trimmed loss (weight=1) that operates on an inverse depth representation [16] and gradient-matching loss [10] (weight=1).”
MethodHADiff: hierarchy aggregated diffusion model for pathology image segmentation
Zhang, X., Yan, B., Xing, Z., Gao, F., Tao, Y., Han, Z., Wang, W., & Zhu, L. (2025) · The Visual Computer · doi:10.1007/s00371-024-03746-z · Cited by 5
“We compare our network with six state-of-the-art methods, including CNN-based methods like UNet [35], UNet++ [77], and DeepLabv3+ [5], as well as Transformer-based approaches like UperNet [56], Seg-former [57], and DPT [33].”
Sankaramaddi, H. R., Lee, W. S., Kim, K.-C., et al. (2026) · Sensors · doi:10.3390/s26061775 · Cited by 4
“For agricultural “digital twins,” this creates metrically stable, mesh-native models of stems, trellises, machinery, and building interiors that work well with CAD/robotics stacks and trait computation downstream [69,76].”
Influential
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
