Pith. sign in

REVIEW 4 major objections 5 minor 53 references

HiResNets: Native Full-HD Video Recognition with Foveal Residual Streams

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read HiResNets embed foveation directly into residual backbones, making convolutional block cost grow logarithmically-squared rather than quadratically with resolution, so full-HD video can be processed natively.

desk verdict Interesting architecture with a real log-squared conv scaling, but the paper overclaims the memory-free construction and under-delivers on empirical support. read the letter →

arxiv 2608.02140 v3 pith:6NF3MAWC submitted 2026-08-03 cs.CV cs.PF

Shivani Mall , Swarnim Jain , Joao F. Henriques This is my paper
classification cs.CVcs.PF
keywords foveatedvisionlog-polarwarpresidualnetworksvideorecognitionegocentrichigh-resolutionefficientinference
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that the quadratic cost of high-resolution video recognition is not inevitable: a residual network whose convolutional blocks interact with a full-resolution buffer only through log-polar warps can keep block compute logarithmic-squared in resolution. The payoff, if true, is that full-HD and higher video can be processed natively, with fine detail preserved in the residual stream while convolutions run on small foveated views. The paper reports that this design scales nearly linearly in latency, learns human-like fixation points, and beats standard baselines on egocentric gaze, object detection, and fine-grained part recognition. A theoretical construction is also given that removes the quadratic memory of the residual stream itself by composing log-polar warps directly.

What carries the argument

The central object is the log-polar warp pair $\psi_c$ and $\psi^{-1}_c$ (Eqs. 3-4 and 7), which maps a Cartesian feature map to a warped view with $\Theta = O(\log W)$ angular and $P = O(\log H)$ radial samples around a predicted focus center $c$. A residual block computes $u = \psi_c(x)$, $y = f(u)$, and $x \leftarrow x + \psi^{-1}_c(y)$, so convolutions run only on the compact log-polar tensor while the full-resolution residual stream acts as a persistent buffer. The paper's complexity argument shows the convolutional cost drops to $O(k^2 \log W \log H)$, and its theoretical construction composes log-polar-to-log-polar warps $W_{k\to l} = \psi_{c_l} \circ \psi^{-1}_{c_k}$ so the stream $x$ is never instantiated.

What would settle it

Measure the reconstruction fidelity of composing the inverse log-polar warp from one focus center with the forward warp from a different center: if $\|\psi_{c_2}(\psi^{-1}_{c_1}(y)) - y\|$ is large for high-frequency peripheral content, the accumulated residual stream has already lost the fine detail the method claims to preserve. Concretely, run the model on a test frame containing small high-contrast targets far from the focus point and compare detection accuracy against the same model with the focus forced onto those targets; a large accuracy drop when targets stay peripheral would indicate the peripheral-sampling assumption fails.

Watch

Extended reading notes

Core claim

The paper claims that residual networks can process full-HD video natively by treating the residual stream as a high-resolution buffer that convolutional blocks read and write only through log-polar warps. Each block's convolutions operate on a small warped view centered on a predicted focus point, so block cost scales as $O(\log W \log H)$ instead of $O(WH)$, and repeated blocks integrate multiple focused updates into the buffer, like saccades building a complete percept. A theoretical construction eliminates the quadratic residual-stream memory by rewriting the accumulated update as sums of direct log-polar-to-log-polar warps $W_{k\to l} = \psi_{c_l} \circ \psi^{-1}_{c_k}$, storing only log-polar tensors. Experiments on egocentric benchmarks report accuracy gains over baselines, especially for small objects and fine-grained part detection, with latency growing nearly linearly rather than quadratically as input resolution increases.

Load-bearing premise

The load-bearing premise is that the bilinear log-polar warp and its inverse are accurate enough that, after many warped updates are accumulated into the residual stream, no task-relevant fine detail is lost; the finite log-polar grid samples the periphery far more coarsely than the fovea, and the paper does not analyze the error of composing the forward and inverse warps.

Editorial extensions

If this is right

  • Residual blocks in HiResNets scale as $O(k^2 \log W \log H)$, so increasing input resolution costs far less than in standard convnets or vision transformers.
  • The residual stream preserves full fidelity, and the theoretical construction removes its quadratic memory by storing only log-polar tensors and composing warps directly, at the price of a quadratic dependency on depth.
  • On egocentric benchmarks (Ego4D, EgoObjects, PACO, EGTEA), the method improves small-object and fine-grained accuracies over YOLOv11, GLC, and SqueezeTime baselines at matched resolutions.
  • Latency grows nearly linearly with input resolution up to 1K, whereas comparable baselines grow quadratically.
  • The learned foveation is interpretable: predicted focus points align with gaze trajectories and alternate between near and far regions in panoramic scenes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: if the warp-composition construction holds as stated, the same design should extend beyond full HD to 4K and 8K inputs, where the memory savings are larger; the paper does not report such an extrapolation.
  • Editorial inference: the center predictor is effectively a differentiable spatial-attention mechanism, so HiResNets could plausibly be combined with token-pruning or sparse-attention transformers to extend log-squared savings beyond convolutional components.
  • Editorial inference: a direct test of the peripheral-sampling assumption would force the focus point away from a small target and measure the accuracy drop; the paper's 'no prediction' ablation already hints that focus location matters, but it does not isolate peripheral reconstruction error.
  • Editorial inference: the theoretical construction shifts cost from resolution to depth, so very deep networks may accumulate warp-composition error; quantifying that drift would show whether log-squared scaling remains practical at scale.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes HiResNets, a residual architecture that applies convolutional blocks in a log-polar warped view around a dynamically predicted focus point, while a full-resolution residual stream is preserved and updated by inverse warps. The authors derive logarithmic-square scaling for the convolutional blocks, present a theoretical construction intended to eliminate the full-resolution residual stream by composing log-polar warps, and report experiments on egocentric gaze estimation, object detection, and fine-grained part detection. The central idea is to embed foveation inside the backbone rather than wrapping a standard network with external glimpse or zoom modules.

Significance. If the log-squared scaling for the convolutional blocks and the warp-composition construction could be made to cover the full adaptive method, the paper would be a useful step toward efficient high-resolution video recognition. The core conceptual contribution is clear: treating the residual stream as a persistent high-resolution buffer that is only read and written through warped views is an elegant way to localize the resolution bottleneck. The complexity derivation is self-contained and does not rely on fitted parameters, which is a strength. However, the paper's empirical support is thin, the appendix that is referenced for key results is absent, and the theoretical construction as written does not apply to the adaptive center predictor. The value of the paper currently lies more in the architecture concept than in the demonstrated results.

major comments (4)
  1. [Sec. 6.3, Eq. (6)] The memory-free construction in Sec. 6.3 eliminates the residual stream only when the centers c_k are fixed or precomputed. In the proposed method, the center predictor receives the current residual state x (Sec. 3.4) and computes a full-resolution attention map a in R^{H×W}; with x never instantiated, the adaptive centers cannot be computed as specified. The paper should either show how c_l is obtained from the stored log-polar tensors {y_k}, or explicitly restrict the construction to fixed centers, which would not represent the proposed adaptive method.
  2. [Sec. 3.6 and Abstract] The claim that the residual stream 'maintains full fidelity' and that 'no fine detail is lost' is not supported. Every update to x is the inverse warp of a log-polar tensor whose peripheral sampling is coarser than the foveal region, and bilinear interpolation cannot reconstruct high-frequency peripheral content that was not sampled at warp time. The paper should provide an error analysis of the warp/unwarp composition, such as a bound on the reconstruction error in the periphery, or soften the fidelity claim to note that the buffer is full-resolution but the updates are lossy outside the fovea.
  3. [Sec. 3.6 and Sec. 5.3] The complexity analysis counts only the convolution in the warped view and omits the O(HW) center predictor, which runs two 1x1 convolutions over a full-resolution attention map, and the O(HW) residual-stream update. Figure 6's caption already concedes that 'linearity dominates,' but the paper still attributes nearly-linear scaling to the whole model. The full pipeline complexity should be reported, and Fig. 5b should separately measure the center predictor, warp, and copy/add contributions to latency.
  4. [Sec. 5.2 and Table 3] The claimed 5 percent accuracy improvement over a ResNet baseline in image classification is not backed by any experiment in the manuscript, and the appendix that is said to contain it is absent. Additionally, Tables 1-3 report no error bars, number of seeds, or training details, and the comparison baselines are not matched in backbone or compute. These omissions make the empirical claim of 'superior performance' hard to verify; add matched-backbone ablations with multiple seeds and error bars.
minor comments (5)
  1. [Sec. 5.1] There is a typo: 'For the the Ego4D dataset' contains a duplicated article.
  2. [Table 1] The column header 'Max Res.' is not defined; clarify whether it refers to training resolution, evaluation resolution, or both.
  3. [Sec. 5.3 and Fig. 5b] The text says latency is measured from 224px to 1K, but Fig. 5b lists resolutions 700, 1.3K, 1.7K, and 2K; the two ranges should be made consistent.
  4. [Sec. 3.2, Eq. (7)] The inverse warp is defined in continuous coordinates but the output grid size is not specified; state explicitly that it maps to the full H by W grid.
  5. [Abstract and Title] The abstract and title claim 'native Full-HD' processing, but the largest input resolution reported in Tables 1-3 is 1400 pixels; no experiment at 1920 by 1080 is shown.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the scaling-law derivation and benchmarks are self-contained, and the only self-citation is non-load-bearing.

full rationale

The paper's central derivation is algebraically self-contained. The log-polar warp sizes Theta = O(log W) and P = O(log H) in Sec. 3.6 follow directly from the warp definition, and the residual-stream elimination identity u_l = sum_{k<l} W_{k->l}(y_k) in Sec. 6.3 follows from the block update equation (5) for fixed centers. No fitted parameter is renamed as a prediction; the accuracy claims are evaluated against held-out Ego4D, EGTEA, EgoObjects, and PACO data with external baselines (GLC, SqueezeTime, YOLOv11). The only author self-citation is [18] (Henriques and Vedaldi), used for the softargmax/STN-style center predictor, which is a standard component and is not load-bearing for the logarithmic-square scaling claim. The paper also explicitly concedes that log-square scaling applies to convolutional blocks only and that residual-stream copy/add and warps remain linear in resolution (Fig. 5b/6 caption), so the complexity claim is not presented as a full-pipeline guarantee. A separate correctness concern, not circularity, is that the Sec. 6.3 construction omits the center predictor that reads x in Eq. 6; eliminating x would remove the input to adaptive foveation, making that extension unsupported rather than a derivation that reduces to its own inputs.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the choice of log-polar grid size and ranges, the differentiable resampling assumption, and a no-error-compounding assumption for the theoretical construction. No new physical entities are introduced.

free parameters (3)
  • Log-polar grid dimensions P and Theta = not specified numerically in the paper
    The claimed O(log W log H) scaling of convolution cost assumes P = O(log H) and Theta = O(log W); the actual grid size is a free design choice. If P and Theta are held constant as resolution grows, the asymptotic claim changes.
  • Residual stream scale factor (input resolution / stream resolution) = 4 (stream is 4x smaller than input)
    A design choice that sets the memory cost of the residual stream; the paper uses a stream 4x smaller than the input, which depends on the stride of the first layer.
  • Radial and angular ranges, rho_min and rho_max = not specified
    The log-polar mapping in Eq. 3 uses exp(rho), so the grid must choose a maximum radius; this is a hand-set hyperparameter affecting warp accuracy and resolution allocation.
assumptions (3)
  • domain assumption The log-polar grid has Theta = O(log W) and P = O(log H) angular and radial samples as input resolution W x H grows.
    Assumed in Sec. 3.6 for the complexity analysis; in practice grid sizes are fixed constants, so the asymptotic claim depends on how the grid is scaled with resolution.
  • standard math Bilinear resampling (Eq. 4) is differentiable and provides a valid gradient signal for the softargmax center predictor.
    Standard practice; not proven in the paper, but widely used in spatial transformer networks.
  • domain assumption The composition of warp psi_c and inverse warp psi^{-1}_c has bounded error, so repeated unwarp-warp cycles do not accumulate destructive artifacts.
    Required for the theoretical construction in Sec. 6.3, where every block recomposes warps W_{k->l}; no error analysis is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HiResNets: Native Full-HD Video Recognition with Foveal Residual Streams." pith.science (2026). https://pith.science/paper/6NF3MAWC

@misc{pith2026260802140,
  author       = {Pith},
  title        = {Pith review of: HiResNets: Native Full-HD Video Recognition with Foveal Residual Streams},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6NF3MAWC}},
  note         = {Machine review of arXiv:2608.02140}
}
read the original abstract

Much of the recent progress in image and video recognition has come at the cost of memory: larger models, increased resolution, and longer temporal contexts. An inevitable component is the quadratic (or larger) growth of memory and compute based on image resolution, which is a property of the grid sampling used in convolutional networks and vision transformers. In this work we study residual networks whose convolutional blocks have logarithmic-square growth instead, enabling them to process very high-resolution video quickly. The key insight is to use a residual architecture's residual stream as a high-resolution buffer, to which convolutional blocks only read and write via log-polar image warp operations. Layers adaptively focus on different parts of each frame, with very high resolution only near the focus point. A complete high-resolution representation is built up in the residual stream, analogous to eye saccades creating a complete picture in biological vision, and a theoretical construction is presented that eliminates the quadratic dependency of the residual stream resolution. Experiments demonstrate that our proposed HiResNets learn to foveate around scenes similarly to human vision, and have superior performance in difficult egocentric video recognition tasks, especially egocentric video with small objects and fine-grained recognition.

Figures

Figures reproduced from arXiv: 2608.02140 by the authors.

Figure 1
Figure 1. (a-c) Illustration of non-uniform resolution on 3 datasets: EgoObjects, PACO and EGTEA. The first row shows original images and focus point (red circles), while the second row shows non-uniform resolution images (by warping and un-warping with a log-polar grid). Notice that regions away from the focus point (in the second row) have much less detail. (d) The same 3 images, warped to log-polar space. The vertical axis… view at source ↗
Figure 2
Figure 2. Overview of the differences between our scheme and alternative multi-gaze strategies. Using multi-gaze crops in every bottleneck [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of the proposed architecture. We integrate [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Qualitative results (randomly selected): [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Overview of timing and qualitative gaze estimation re [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Comparison of computational efficiency and detection [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Review history (2 revisions) →

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

53 extracted references · 49 canonical work pages

  1. [1]

    Salisa: Saliency-based input sampling for efficient video object detection, 2022

    Babak Ehteshami Bejnordi, Amirhossein Habibian, Fatih Porikli, and Amir Ghodrati. Salisa: Saliency-based input sampling for efficient video object detection, 2022. 8

    work page 2022

  2. [2]

    Token merging: Your vit but faster

    Daniel Bolya, Cheng-Yang Fu, Xiaoliang Dai, Peizhao Zhang, Christoph Feichtenhofer, and Judy Hoffman. Token merging: Your vit but faster. InInternational Conference on Learning Representations (ICLR), 2023. 4

    work page 2023

  3. [3]

    Visual attention: The past 25 years.Vision Research, 51(13):1484–1525, 2011

    Marisa Carrasco. Visual attention: The past 25 years.Vision Research, 51(13):1484–1525, 2011. 2

    work page 2011

  4. [4]

    Chen, Chengkuan Chen, Yicong Li, Tiffany Y

    Richard J. Chen, Chengkuan Chen, Yicong Li, Tiffany Y . Chen, Andrew D. Trister, Rahul G. Krishnan, and Faisal Mahmood. Scaling vision transformers to gigapixel images via hierarchical self-supervised learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 4

    work page 2022

  5. [5]

    Training deep nets with sublinear memory cost, 2016

    Tianqi Chen, Bing Xu, Chiyuan Zhang, and Carlos Guestrin. Training deep nets with sublinear memory cost, 2016. 1

    work page 2016

  6. [6]

    Dynamic convolution: At- tention over convolution kernels

    Yinpeng Chen, Xiyang Dai, Mengchen Liu, Dongdong Chen, Lu Yuan, and Zicheng Liu. Dynamic convolution: At- tention over convolution kernels. InCVPR, 2020. 7, 8

    work page 2020

  7. [7]

    Curcio, Kenneth R

    Christine A. Curcio, Kenneth R. Sloan, Richard E. Kalina, and Alan E. Hendrickson. Human photoreceptor topography. Journal of Comparative Neurology, 292(4):497–523, 1990. 2

    work page 1990

  8. [8]

    Deformable convolutional networks

    Jifeng Dai, Haozhi Qi, Yuwen Xiong, Yi Li, Guodong Zhang, Han Hu, and Yichen Wei. Deformable convolutional networks. InICCV, 2017. 7, 8

    work page 2017

Show all 53 references
  1. [9]

    Scaling egocentric vision: The EPIC- KITCHENS dataset

    Dima Damen, Hazel Doughty, Giovanni Maria Farinella, Sanja Fidler, Antonino Furnari, Evangelos Kazakos, Da- vide Moltisanti, Jonathan Munro, Toby Perrett, Will Price, and Michael Wray. Scaling egocentric vision: The EPIC- KITCHENS dataset. InEuropean Conference on Computer Vis...

    2018

  2. [10]

    An image is worth 16x16 words: Transformers for image recognition at scale

    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 ...

    2021

  3. [11]

    Saccader: Improving accuracy of hard attention models for vision

    Gamaleldin F Elsayed, Simon Kornblith, and Quoc V Le. Saccader: Improving accuracy of hard attention models for vision. InAdvances in Neural Information Processing Sys- tems, 2019. 1, 3

    2019

  4. [12]

    Polar transformer networks

    Carlos Esteves, Christine Allen-Blanchette, Xiaowei Zhou, and Kostas Daniilidis. Polar transformer networks. InInter- national Conference on Learning Representations, 2018. 3, 4

    2018

  5. [13]

    Multiscale vision transformers

    Haoqi Fan, Bo Xiong, Karttikeya Mangalam, Yanghao Li, Zhicheng Yan, Jitendra Malik, and Christoph Feichten- hofer. Multiscale vision transformers. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2021. 4

    2021

  6. [14]

    Morariu, and Larry S

    Mingfei Gao, Ruichi Yu, Ang Li, Vlad I. Morariu, and Larry S. Davis. Dynamic zoom-in network for fast object detection in large images. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018. 3

    2018

  7. [15]

    Garbin, Yiru Shen, Dushyant Goodman, Jakob H

    Stephan J. Garbin, Yiru Shen, Dushyant Goodman, Jakob H. Lagergren, and Sachin S. Talathi. OpenEDS: Open eye dataset.arXiv preprint arXiv:1905.03702, 2019. 3

    1905 arXiv

  8. [16]

    Ego4D: Around the world in 3,000 hours of egocentric video

    Kristen Grauman et al. Ego4D: Around the world in 3,000 hours of egocentric video. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 3, 6, 7

    2022

  9. [17]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. InProceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 1, 4, 5, 6

    2016

  10. [18]

    Warped convolutions: Efficient invariance to spatial transformations

    Joao F Henriques and Andrea Vedaldi. Warped convolutions: Efficient invariance to spatial transformations. InInterna- tional Conference on Machine Learning, pages 1461–1469. PMLR, 2017. 5, 7, 8

    2017

  11. [19]

    Spatial transformer networks

    Max Jaderberg, Karen Simonyan, Andrew Zisserman, and Koray Kavukcuoglu. Spatial transformer networks. InAd- vances in Neural Information Processing Systems, 2015. 2, 7, 8

    2015

  12. [20]

    Ultralytics yolo11, 2024

    Glenn Jocher and Jing Qiu. Ultralytics yolo11, 2024. 5, 7, 8

    2024

  13. [21]

    Manjunath, and Miguel P

    Aditya Jonnalagadda, William Yang Wang, B.S. Manjunath, and Miguel P. Eckstein. FoveaTer: Foveated transformer for image classification.arXiv preprint arXiv:2105.14173,

    arXiv

  14. [22]

    Gaze360: Physically uncon- strained gaze estimation in the wild

    Petr Kellnhofer, Adri `a Recasens, Simon Stent, Wojciech Ma- tusik, and Antonio Torralba. Gaze360: Physically uncon- strained gaze estimation in the wild. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2019. 3

    2019

  15. [23]

    Paul Siebert

    George Killick, Gerardo Aragon-Camarasa, and J. Paul Siebert. Monte-carlo convolutions on foveated images. In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP) – VISAPP, 2022. 3

    2022

  16. [24]

    Foveation in the era of deep learn- ing

    George Killick, Paul Henderson, Jan Paul Siebert, and Ger- ardo Aragon-Camarasa. Foveation in the era of deep learn- ing. InBritish Machine Vision Conference (BMVC), 2023. 1, 3

    2023

  17. [25]

    In the eye of transformer: Global–local correlation for egocentric gaze estimation and beyond.International Journal of Com- puter Vision, pages 1–18, 2023

    Bolin Lai, Miao Liu, Fiona Ryan, and James M Rehg. In the eye of transformer: Global–local correlation for egocentric gaze estimation and beyond.International Journal of Com- puter Vision, pages 1–18, 2023. 5, 7

    2023

  18. [26]

    Learning to combine foveal glimpses with a third-order boltzmann ma- chine

    Hugo Larochelle and Geoffrey E Hinton. Learning to combine foveal glimpses with a third-order boltzmann ma- chine. InAdvances in Neural Information Processing Sys- tems, 2010. 1, 3

    2010

  19. [27]

    Froehlich, Yuhang Zhao, and Yapeng Tian

    Jia Li, Wenjie Zhao, Shijian Deng, Bolin Lai, Yuheng Wu, Ruijia Chen, Jon E. Froehlich, Yuhang Zhao, and Yapeng Tian. Autoregressive transformers for online egocentric gaze estimation, 2026. 7

    2026

  20. [28]

    Yin Li, Miao Liu, and James M. Rehg. In the eye of the beholder: Joint learning of gaze and actions in first person video. InEuropean Conference on Computer Vision (ECCV),

  21. [29]

    MViTv2: Improved multiscale vision transformers for classification and detection

    Yanghao Li, Chao-Yuan Wu, Haoqi Fan, Karttikeya Man- galam, Bo Xiong, Jitendra Malik, and Christoph Feichten- hofer. MViTv2: Improved multiscale vision transformers for classification and detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogn...

    2022

  22. [30]

    Evit: Expediting vision transform- ers via token reorganizations

    Youwei Liang, Chongjian Ge, Zhan Tong, Yibing Song, Jue Wang, and Pengtao Xie. Evit: Expediting vision transform- ers via token reorganizations. InInternational Conference on Learning Representations (ICLR), 2022. 4

    2022

  23. [31]

    Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie

    Tsung-Yi Lin, Piotr Doll ´ar, Ross B. Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyra- mid networks for object detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2017. 3

    2017

  24. [32]

    Girshick, Kaiming He, and Piotr Doll ´ar

    Tsung-Yi Lin, Priya Goyal, Ross B. Girshick, Kaiming He, and Piotr Doll ´ar. Focal loss for dense object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2017. 3

    2017

  25. [33]

    A convnet for the 2020s

    Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feicht- enhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s. InProceedings of the IEEE conference on computer vision and pattern recognition, 2022. 1

    2022

  26. [34]

    Lu, Drew F.K

    Ming Y . Lu, Drew F.K. Williamson, Tiffany Y . Chen, Richard J. Chen, Matteo Barbieri, and Faisal Mahmood. Data-efficient and weakly supervised computational pathol- ogy on whole-slide images.Nature Biomedical Engineering, 5:555–570, 2021. 4

    2021

  27. [35]

    Kernel foveated rendering.Proceedings of the ACM on Computer Graphics and Interactive Techniques, 1: 1–20, 2018

    Xiaoxu Meng, Ruofei Du, Matthias Zwicker, and Amitabh Varshney. Kernel foveated rendering.Proceedings of the ACM on Computer Graphics and Interactive Techniques, 1: 1–20, 2018. 1, 2

    2018

  28. [36]

    Recurrent models of visual attention

    V olodymyr Mnih, Nicolas Heess, Alex Graves, and Koray Kavukcuoglu. Recurrent models of visual attention. InAd- vances in Neural Information Processing Systems, 2014. 1, 3

    2014

  29. [37]

    Emergence of Fixational and Saccadic Movements in a Multi-level Recurrent Attention Model for Vision, page 299–313

    Pengcheng Pan, Shogo Yonekura, and Yasuo Kuniyoshi. Emergence of Fixational and Saccadic Movements in a Multi-level Recurrent Attention Model for Vision, page 299–313. Springer Nature Singapore, 2025. 1, 3

    2025

  30. [38]

    HD-EPIC: A highly-detailed egocentric video dataset

    Toby Perrett, Ahmad Darkhalil, Saptarshi Sinha, Omar Emara, Sam Pollard, Kranti Parida, Kaiting Liu, Prajwal Gatti, Siddhant Bansal, Kevin Flanagan, Jacob Chalk, Zhi- fan Zhu, Rhodri Guerrier, Fahd Abdelazim, Bin Zhu, Da- vide Moltisanti, Michael Wray, Hazel Doughty, and Dima ...

    2025

  31. [39]

    Paco: Parts and attributes of common objects

    Vignesh Ramanathan, Anmol Kalia, Vladan Petrovic, Yi Wen, Baixue Zheng, Baishan Guo, Rui Wang, Aaron Mar- quez, Rama Kovvuri, Abhishek Kadian, Amir Mousavi, Yi- wen Song, Abhimanyu Dubey, and Dhruv Mahajan. Paco: Parts and attributes of common objects. InProceedings of the IEE...

    2023

  32. [40]

    Dynamicvit: Efficient vision transformers with dynamic token sparsification

    Yongming Rao, Wenliang Zhao, Benlin Liu, Jiwen Lu, Jie Zhou, and Cho-Jui Hsieh. Dynamicvit: Efficient vision transformers with dynamic token sparsification. InAdvances in Neural Information Processing Systems, 2021. 4

    2021

  33. [41]

    Learning to zoom: a saliency- based sampling layer for neural networks

    Adri `a Recasens, Petr Kellnhofer, Simon Stent, Wojciech Ma- tusik, and Antonio Torralba. Learning to zoom: a saliency- based sampling layer for neural networks. InEuropean Con- ference on Computer Vision (ECCV), 2018. 7, 8

    2018

  34. [42]

    Balas, and Livia Ilie

    Ruth Rosenholtz, Jie Huang, Alvin Raj, Benjamin J. Balas, and Livia Ilie. A summary statistic representation in periph- eral vision explains visual search.Journal of Vision, 12(4): 14–14, 2012. 2

    2012

  35. [43]

    Schwartz

    Eric L. Schwartz. Spatial mapping in the primate sensory projection: Analytic structure and relevance to perception. Biological Cybernetics, 25(4):181–194, 1977. 2

    1977

  36. [44]

    Computational anatomy and functional ar- chitecture of striate cortex: A spatial mapping approach to perceptual coding.Vision Research, 20(8):645–669, 1980

    Eric L Schwartz. Computational anatomy and functional ar- chitecture of striate cortex: A spatial mapping approach to perceptual coding.Vision Research, 20(8):645–669, 1980. 1

    1980

  37. [45]

    Pe- ripheral vision and pattern recognition: A review.Journal of Vision, 11(5):13–13, 2011

    Hans Strasburger, Ingo Rentschler, and Martin J ¨uttner. Pe- ripheral vision and pattern recognition: A review.Journal of Vision, 11(5):13–13, 2011. 2

    2011

  38. [46]

    Log-polar space convolution lay- ers

    Bing Su and Ji-Rong Wen. Log-polar space convolution lay- ers. InAdvances in Neural Information Processing Systems,

  39. [47]

    Deep high-resolution representation learning for human pose es- timation

    Ke Sun, Bin Xiao, Dong Liu, and Jingdong Wang. Deep high-resolution representation learning for human pose es- timation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 4

    2019

  40. [48]

    Fovea: Foveated image magnification for autonomous navigation

    Chittesh Thavamani, Mengtian Li, Nicolas Cebron, and Deva Ramanan. Fovea: Foveated image magnification for autonomous navigation. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 15519–15528, 2021. 8

    2021

  41. [49]

    Deep high-resolution representation learning for visual recogni- tion.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020

    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.IEEE Transactions on Pattern Analysis and Machine Intell...

    2020

  42. [50]

    Zoom-in-net: Deep mining lesions for diabetic retinopathy detection

    Zhe Wang, Yanxin Yin, Jianping Shi, Wei Fang, Hongsheng Li, and Xiaogang Wang. Zoom-in-net: Deep mining lesions for diabetic retinopathy detection. InMedical Image Com- puting and Computer Assisted Intervention – MICCAI 2017, pages 267–275, 2017. 1, 3

    2017

  43. [51]

    No time to waste: Squeeze time into channel for mobile video understanding, 2024

    Yingjie Zhai, Wenshuo Li, Yehui Tang, Xinghao Chen, and Yunhe Wang. No time to waste: Squeeze time into channel for mobile video understanding, 2024. 5, 7

    2024

  44. [52]

    MPIIGaze: Real-world dataset and deep appearance-based gaze estimation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(1):162–175,

    Xucong Zhang, Yusuke Sugano, Mario Fritz, and An- dreas Bulling. MPIIGaze: Real-world dataset and deep appearance-based gaze estimation.IEEE Transactions on Pattern Analysis and Machine Intelligence, 41(1):162–175,

  45. [53]

    ETH-XGaze: A large scale dataset for gaze estimation under extreme head pose and gaze variation

    Xucong Zhang, Seonwook Park, Thabo Beeler, Derek Bradley, Siyu Tang, and Otmar Hilliges. ETH-XGaze: A large scale dataset for gaze estimation under extreme head pose and gaze variation. InEuropean Conference on Com- puter Vision (ECCV), 2020. 3

    2020

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.