Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Exploring Temporally-Aware Features for Point Tracking

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Chrono shows that a frozen DINOv2 backbone fitted with inter-block temporal adapters can track arbitrary points by simple feature matching, with no refinement stage, and outperforms standard point-tracking backbones on TAP-Vid.

desk verdict Useful empirical result, but the paper's central causal claim about temporal awareness is untested because the DINOv2 baseline is zero-shot and no N=1 control exists. read the letter →

arxiv 2501.12218 v2 pith:XNNBRONM submitted 2025-01-21 cs.CV

Inès Hyeonsu Kim , Seokju Cho , Jiahui Huang , Jung Yi , Joon-Young Lee , Seungryong Kim This is my paper · ORCID
classification cs.CV
keywords pointtrackingtemporaladapterDINOv2TAP-Vidvideocorrespondencefeaturebackbonesoft-argmaxrefiner-free
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

Point tracking usually needs a two-stage pipeline: a coarse backbone predicts rough positions, then a heavy refiner injects temporal context and corrects errors. The paper argues this division is unnecessary if the backbone itself is temporally aware. It introduces Chrono, which freezes DINOv2 and inserts small temporal adapters between its transformer blocks, then tracks points by plain soft-argmax feature matching with no learnable prediction layers. On the TAP-Vid benchmarks, Chrono's backbone-only tracks beat the common ResNet, TSM-ResNet, CoTracker, and DINOv2 backbones, and match or approach full refiner-based trackers at much higher throughput.

What carries the argument

The carrying object is the temporal adapter, a bottleneck layer inserted between each frozen DINOv2 transformer block. It compresses spatial resolution by stride 4, applies 1D local attention across a window of 13 frames at each spatial location, restores resolution, and adds a residual connection back to the frozen features. The 1D attention lets the network weight which neighboring frames matter for each query, and the residual preserves DINOv2's pretrained spatial knowledge. Tracking itself is a non-parametric operation: a correlation map via cosine similarity plus a masked soft argmax. The adapter placement ablation shows that putting the adapter between all 12 blocks gives the best accuracy, and the aggregation ablation shows that 1D attention beats 1D and 3D convolutions.

What would settle it

Fine-tune the frozen DINOv2 backbone on Kubric with the same Huber loss, soft-argmax protocol, and training iterations but no temporal adapters, then evaluate on TAP-Vid-DAVIS strided; if that control reaches roughly the same $<\delta^x_{\mathrm{avg}}$ as Chrono, the temporal adapter is not the cause of the improvement.

Watch

Extended reading notes

Core claim

The central claim is that temporal awareness belongs inside the feature backbone, not in a post-hoc refiner. Chrono combines frozen DINOv2 spatial features with inter-block temporal adapters, each doing 2D convolutional downsampling, 1D local attention over a temporal window of 13 frames, and a residual connection. Using only cosine-similarity matching and a masked soft argmax, without any learnable layer after feature extraction, Chrono (ViT-S/14) reaches 68.0 $<\delta^x_{\mathrm{avg}}$ on TAP-Vid-DAVIS strided versus 50.4 for DINOv2 (ViT-S/14) and 53.3 for ResNet-18; Chrono (ViT-B/14) reaches 70.1. The paper concludes that refiner-free tracking with temporally-aware pre-trained features is both more accurate and substantially faster, with 12.5x the throughput of TAPIR on this setting.

Load-bearing premise

The paper's central comparison is not controlled: Chrono was trained with a tracking loss on synthetic data, while the DINOv2 baseline was evaluated zero-shot, so supervised adaptation alone could account for much of the reported gain.

Editorial extensions

If this is right

  • Chrono's backbone-only tracks exceed 90% at the $<\delta_4$ threshold on TAP-Vid-DAVIS strided, meaning subpixel-level precision is achievable without a refinement stage.
  • Dropping Chrono into LocoTrack's refiner improves Average Jaccard on all TAP-Vid datasets, with the largest gain of +6.1 on RGB-Stacking, so existing pipelines can swap in this backbone.
  • Throughput of 26k to 34k points per second, versus about 2k for TAPIR, suggests the temporal-adapter design removes the per-query refinement bottleneck for real-time and robotics use.
  • Because the temporal window of 13 frames gives six times more temporal context than TSM-ResNet, the result points to longer-range temporal context as a key ingredient for accurate point tracking.

Reading between the lines

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

  • Editorial inference: the comparison that would isolate temporal awareness, fine-tuning DINOv2 with identical training but no temporal adapter, is absent from the paper, so the causal role of the adapter should be treated as plausible but unproven.
  • Editorial inference: Chrono's soft-argmax prediction operates on a single-resolution feature map, so a multi-scale variant might close the remaining gap to refiner-based trackers on Kinetics, where motion blur and abrupt cuts are frequent.
  • Editorial inference: because the adapter operates at each spatial location independently, the same frozen-backbone-plus-temporal-adapter recipe could transfer to other video correspondence tasks such as optical flow, dense tracking, or video object segmentation.
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 Chrono, a point-tracking feature backbone built from frozen DINOv2 with inter-block temporal adapters that mix information across frames via 2D convolutions and 1D local temporal attention. Tracks are produced by correlating a query feature with target-frame features and applying soft-argmax, with no additional learned refinement. The authors report that Chrono substantially outperforms existing point-tracking backbones (ResNet-18, TSM-ResNet-18, CoTracker backbone) and zero-shot DINOv2 on TAP-Vid-RGB-Stacking, TAP-Vid-Kinetics, and TAP-Vid-DAVIS, in both strided and first-query modes, while being faster than refiner-based pipelines. They also show that Chrono can be combined with LocoTrack's refiner to further improve state-of-the-art results, and they ablate the temporal aggregation design and adapter placement. The central claim is that embedding temporal awareness directly in the backbone enables accurate tracking without a refinement stage.

Significance. If the central claim is established, the paper makes a useful contribution by showing that a pre-trained, temporally adapted backbone can replace the expensive per-query refinement stage in point tracking. The empirical study is broad: it covers multiple TAP-Vid datasets, two query modes, comparisons with several backbones and full pipelines, an integration with LocoTrack, and ablations of the adapter design and placement. The efficiency numbers (Tables 3 and 7) and the finding that a simple non-parametric soft-argmax head suffices are practically important. The paper also clearly identifies a previously underexplored design axis: temporal awareness inside the backbone rather than in a separate refiner. The reported gains are large, and the ablations suggest that adapter placement and temporal aggregation type matter. However, as detailed below, the key causal attribution of these gains to temporal awareness is not yet supported by the experiments as presented.

major comments (4)
  1. [Section 4.3, Tables 1 and 2] The main comparison against DINOv2 is not controlled: Chrono is trained on Kubric with a tracking loss and soft-argmax supervision, while DINOv2 is evaluated zero-shot with no adaptation. The large DAVIS gap (e.g., <δx_avg 68.0 for Chrono ViT-S/14 vs. 50.4 for DINOv2 ViT-S/14 in Table 1) therefore conflates the effect of temporal awareness with the effects of supervised adaptation to the tracking task and the added trainable capacity of the adapters. To support the paper's central claim that temporal awareness is what drives the improvement, the authors need an additional control: a DINOv2 backbone fine-tuned on Kubric with the same Huber loss and soft-argmax protocol but with the temporal adapter removed (or with the temporal window set to N=1), keeping all other training details identical. Without this control, the abstract's attribution of the gain to temporal awareness is not established.
  2. [Section 4.4, Tables 5 and 6] The ablations vary the temporal aggregation operation (1D Conv, 3D Conv, 1D Attn) and the placement of adapters, but every variant retains cross-frame temporal mixing. There is no ablation with the temporal window effectively disabled (N=1) or with temporal attention removed while keeping the same trainable adapter capacity and training protocol. Such an ablation is necessary to determine how much of the improvement over zero-shot DINOv2 comes from the temporal mechanism itself and how much comes from simply adding trainable layers and supervised training to a frozen DINOv2. The current experiments cannot rule out that a non-temporal adapter (e.g., a spatial-only conv adapter) would achieve similar gains.
  3. [Section 3.2 and Figure 4] Figure 4 labels the projection layers within the temporal adapter as 'LoRA Adapter,' but Section 3.2 never defines LoRA, its rank, or how low-rank adaptation is applied to W_Q, W_K, and W_V. Since the trainable parameterization is a core architectural element, the manuscript should specify the LoRA configuration (rank, alpha, dropout) and clarify whether the adapter's linear projections use LoRA or full-rank weights, as this affects both the parameter count reported in Table 3 and the reproducibility of the method.
  4. [Section 4.3, Tables 1 and 2] No variance or error bars are reported for any of the benchmark numbers. Because the TAP-Vid evaluation sets are small (30 videos for DAVIS, 50 for RGB-Stacking) and the differences between some configurations (e.g., DINOv2 ViT-B vs. ViT-S in Table 1, or Chrono ViT-S vs. ViT-B in Kinetics) are a few points, the authors should either report uncertainty estimates (e.g., bootstrap over videos) or state that the differences are within noise. This is especially important for the comparison between Chrono and the CoTracker backbone on Kinetics, where the gap is small.
minor comments (5)
  1. [Abstract] The sentence 'These approach, however, is computationally expensive...' contains a subject-verb agreement error; it should be 'This approach' or 'These approaches.'
  2. [Section 4.3, Table 1] The name 'CoTraker [25] Backbone' is misspelled; it should be 'CoTracker.' Also, the table caption says 'Best scores are in bold and second best are underlined,' but the table does not visibly show underlining in the text; please check the rendering.
  3. [Section 4.3, Table 3] The header 'Additional Params.' is ambiguous: it lists 0M for Chrono, but the model has 16.2M trainable parameters in the adapters (as stated in the table and Section 3.2). The column should be labeled 'Refiner Params.' or 'Trainable Params.' to avoid confusion, and the text should clarify which parameter count is being reported.
  4. [Section 4.2] The definitions of the evaluation metrics (<δ0, <δ1, etc.) are clear, but the text says 'position accuracy at five threshold levels' with thresholds of 1, 2, 4, 8, and 16 pixels. Please double-check the notation: <δ0 corresponds to 1 pixel, <δ1 to 2 pixels, etc., which is nonstandard and could be clarified in a parenthetical.
  5. [Section 4.4, Table 7] The table reports inference time for feature extraction without point prediction, but the main efficiency claim in Table 3 compares full tracking throughput. Please clarify in the text what the numbers in Table 7 include (e.g., GPU warm-up, batching) and whether the reported Chrono times include the temporal adapter overhead only or also the soft-argmax prediction.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Chrono's benchmark gains are empirical and evaluated against external baselines; missing controls are experimental-design concerns, not circular reductions.

full rationale

The paper makes an empirical architecture claim: a DINOv2 backbone augmented with inter-block temporal adapters, trained with a Huber loss on Kubric, is evaluated on the external TAP-Vid benchmark against published baselines. The reported gains are not derived from a fitted parameter renamed as a prediction, nor does any equation reduce a claimed output to an input. The soft-argmax predictor (Eq. 6-8) is a fixed, non-learnable operation applied to the learned features, and the temporal adapter (Eq. 1-5) is a concrete architectural module whose design is ablated in Tables 5 and 6. Hyperparameters such as tau, M, and N are hand-chosen and reported, but they do not make the evaluation circular. The self-citations (LocoTrack for the Kubric Panning-MOVi-E training split, FlowTrack as related work, and LocoTrack as a refiner in Table 4) are not load-bearing: the central claim does not rest on an unverified assertion from the authors' prior work, and the Table 4 integration test is an empirical comparison rather than a derivation. The skeptic's concerns are real experimental-design gaps, not circularity: the DINOv2 baseline in Tables 1 and 2 is evaluated zero-shot without supervised adaptation, so the temporal-awareness attribution is confounded by training; Table 5 varies the temporal aggregation layer but never disables temporal mixing entirely; and Figure 4 mentions a 'LoRA Adapter' that is never defined in Section 3.2. These are missing-control and reproducibility issues that weaken the causal claim, but they do not constitute a reduction of the result to its inputs. No circular step can be quoted from the paper, so per the hard rules the score is 0.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The paper is an empirical architecture study; its central claim rests on supervised training on a synthetic dataset, transfer to a real benchmark, and several hand-chosen hyperparameters. No new physical or mathematical entities are postulated.

free parameters (5)
  • Softmax temperature tau = 20.0
    Chosen by hand; controls sharpness of soft-argmax and affects position accuracy.
  • Soft-argmax mask radius M = 5 pixels
    Positions farther than 5 pixels from the correlation maximum are masked out; chosen by hand.
  • Temporal window size N = 13 frames
    Defines the local temporal attention window; chosen to balance efficiency and accuracy.
  • Conv2D stride s in temporal adapter = 4
    Spatial downsampling factor in each adapter; chosen by hand.
  • Number of temporal adapters = 11 (all inter-block positions)
    Ablation shows all-blocks works best; the count is a design choice.
assumptions (4)
  • domain assumption DINOv2 features, trained on large-scale real images, provide a suitable spatial representation for point matching.
    The whole method relies on frozen DINOv2 features remaining useful after adapter insertion.
  • domain assumption Training on Kubric Panning MOVi-E synthetic data transfers to real TAP-Vid videos.
    The model is trained only on synthetic data and evaluated on real videos; no real training data is used.
  • domain assumption A local temporal window of 13 frames is sufficient to capture the motion dynamics needed for tracking.
    The temporal adapter attends only to frames within 6 frames of the current one; longer-range dependencies are not directly modeled.
  • domain assumption Occluded points can be excluded from the loss without biasing the learned features.
    Loss masking of occluded points assumes the model need not predict through occlusions during training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exploring Temporally-Aware Features for Point Tracking." pith.science (2026). https://pith.science/paper/XNNBRONM

@misc{pith2026250112218,
  author       = {Pith},
  title        = {Pith review of: Exploring Temporally-Aware Features for Point Tracking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XNNBRONM}},
  note         = {Machine review of arXiv:2501.12218}
}
read the original abstract

Point tracking in videos is a fundamental task with applications in robotics, video editing, and more. While many vision tasks benefit from pre-trained feature backbones to improve generalizability, point tracking has primarily relied on simpler backbones trained from scratch on synthetic data, which may limit robustness in real-world scenarios. Additionally, point tracking requires temporal awareness to ensure coherence across frames, but using temporally-aware features is still underexplored. Most current methods often employ a two-stage process: an initial coarse prediction followed by a refinement stage to inject temporal information and correct errors from the coarse stage. These approach, however, is computationally expensive and potentially redundant if the feature backbone itself captures sufficient temporal information. In this work, we introduce Chrono, a feature backbone specifically designed for point tracking with built-in temporal awareness. Leveraging pre-trained representations from self-supervised learner DINOv2 and enhanced with a temporal adapter, Chrono effectively captures long-term temporal context, enabling precise prediction even without the refinement stage. Experimental results demonstrate that Chrono achieves state-of-the-art performance in a refiner-free setting on the TAP-Vid-DAVIS and TAP-Vid-Kinetics datasets, among common feature backbones used in point tracking as well as DINOv2, with exceptional efficiency. Project page: https://cvlab-kaist.github.io/Chrono/

Figures

Figures reproduced from arXiv: 2501.12218 by the authors.

Figure 1
Figure 1. Chrono is a highly precise, temporally-aware feature backbone specifically designed for point tracking. Unlike prior methods that rely on heavy iterative refinement for each query point [11–13, 19, 25] or test-time optimization [46, 49], Chrono achieves competitive tracking performance through simple feature matching using a non-parametric operation. Moreover, its perfor￾mance can be further improved by an optional … view at source ↗
Figure 2
Figure 2. Comparison of initial trajectories from Chrono and existing feature backbone for point tracking. Chrono demon￾strates a significant improvement over existing feature backbones such as TSM-ResNet [32]. tion [11, 14, 25, 26, 31, 48]. Therefore, features used in point tracking also need to be temporally aware. Despite its critical role, temporally-aware backbone has been relatively under-explored in this area. Early ef… view at source ↗
Figure 3
Figure 3. Overall architecture of Chrono. Temporal adapters be￾tween transformer blocks use 2D convolution and 1D local atten￾tion along the temporal axis to output temporally-aware features. given a video sequence {It} T −1 t=0 , where It ∈ R H×W×3 represents the t-th frame of height H and width W, and a query point q = (xq, yq, tq) ∈ R 3 at frame tq, the goal is to produce a trajectory T = {pˆt} T −1 t=0 , where pˆt ∈ R 2 d… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Point track prediction. Feed-Forward To predict point positions, we simply match the query points with features from other frames, without using any learnable layers. the value vectors from these neighboring frames: f attn t (x, y) = X k n=−k α (t,n) (x, y) · Vt+n(x, y…
Figure 5
Figure 5. Figure 5: Qualitative comparison of complex real-world video tracking. We qualitatively compare the results generated by Chrono with those from other commonly used backbones in point tracking and as well as DINOv2. Our model demonstrates better smoothness and precision than othe…
Figure 6
Figure 6. Figure 6: Visualization of features. We visualize the features generated by our model and DINOv2 using PCA. The results show that our model demonstrates improved temporal smoothness and finer-grained feature representation relative to DINOv2. The results suggest that 1D attentio…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Improving Video Diffusion Transformer Training by Multi-Feature Fusion and Alignment from Self-Supervised Vision Encoders

    cs.CV 2025-09 conditional novelty 5.0 of 10

    Matching video diffusion transformer tokens to concatenated DINOv2 and SAM2 features improves FVD/FID and speeds convergence, e.g., 400K-step fusion beats 1M-step baseline on UCF-101.

Reference graph

Works this paper leans on

54 extracted references · 25 canonical work pages · cited by 1 Pith paper

  1. [1]

    Deep vit features as dense visual descriptors

    Shir Amir, Yossi Gandelsman, Shai Bagon, and Tali Dekel. Deep vit features as dense visual descriptors. arXiv preprint arXiv:2112.05814, 2(3):4, 2021. 3

    arXiv 2021

  2. [2]

    Can Visual Foundation Models Achieve Long-term Point Tracking?

    G ¨orkay Aydemir, Weidi Xie, and Fatma G ¨uney. Can visual foundation models achieve long-term point tracking? arXiv preprint arXiv:2408.13575, 2024. 3

    work page Pith review arXiv 2024

  3. [3]

    Long- former: The long-document transformer

    Iz Beltagy, Matthew E Peters, and Arman Cohan. Long- former: The long-document transformer. arXiv preprint arXiv:2004.05150, 2020. 4

    arXiv 2004

  4. [4]

    Zoedepth: Zero-shot trans- fer by combining relative and metric depth

    Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, and Matthias M ¨uller. Zoedepth: Zero-shot trans- fer by combining relative and metric depth. arXiv preprint arXiv:2302.12288, 2023. 1

    arXiv 2023

  5. [5]

    Depth pro: Sharp monocular metric depth in less than a second

    Aleksei Bochkovskii, Ama ¨el Delaunoy, Hugo Germain, Marcel Santos, Yichao Zhou, Stephan R Richter, and Vladlen Koltun. Depth pro: Sharp monocular metric depth in less than a second. arXiv preprint arXiv:2410.02073, 2024. 4

    arXiv 2024

  6. [6]

    End-to- end object detection with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to- end object detection with transformers. In European confer- ence on computer vision, pages 213–229. Springer, 2020. 1, 3

    2020

  7. [7]

    Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolu- tion, and fully connected crfs

    Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolu- tion, and fully connected crfs. IEEE transactions on pattern analysis and machine intelligence, 40(4):834–848, 2017

    work page 2017

  8. [8]

    Schwing, Alexan- der Kirillov, and Rohit Girdhar

    Bowen Cheng, Ishan Misra, Alexander G. Schwing, Alexan- der Kirillov, and Rohit Girdhar. Masked-attention mask transformer for universal image segmentation. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 1290–1299, 2022. 1

    work page 2022

Show all 54 references
  1. [9]

    Cats: Cost ag- gregation transformers for visual correspondence

    Seokju Cho, Sunghwan Hong, Sangryul Jeon, Yunsung Lee, Kwanghoon Sohn, and Seungryong Kim. Cats: Cost ag- gregation transformers for visual correspondence. Advances in Neural Information Processing Systems , 34:9011–9023,

  2. [10]

    Flowtrack: Revisiting optical flow for long- range dense tracking

    Seokju Cho, Jiahui Huang, Seungryong Kim, and Joon- Young Lee. Flowtrack: Revisiting optical flow for long- range dense tracking. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19268–19277, 2024. 2, 3, 4, 7

    2024

  3. [11]

    Local all-pair correspon- dence for point tracking

    Seokju Cho, Jiahui Huang, Jisu Nam, Honggyu An, Seun- gryong Kim, and Joon-Young Lee. Local all-pair correspon- dence for point tracking. arXiv preprint arXiv:2407.15420,

    arXiv

  4. [12]

    Tap-vid: A benchmark for track- ing any point in a video

    Carl Doersch, Ankush Gupta, Larisa Markeeva, Adria Re- casens, Lucas Smaira, Yusuf Aytar, Joao Carreira, Andrew Zisserman, and Yi Yang. Tap-vid: A benchmark for track- ing any point in a video. Advances in Neural Information Processing Systems, 35:13610–13626, 2022. 2, 3, 5, 6, 7

    2022

  5. [13]

    TAPIR: Tracking any point with per-frame initialization and temporal refinement

    Carl Doersch, Yi Yang, Mel Vecerik, Dilara Gokay, Ankush Gupta, Yusuf Aytar, Joao Carreira, and Andrew Zisserman. TAPIR: Tracking any point with per-frame initialization and temporal refinement. In Proceedings of the IEEE/CVF In- ternational Conference on Computer Vision , pag...

    2023

  6. [14]

    Bootstap: Boot- strapped training for tracking-any-point

    Carl Doersch, Pauline Luc, Yi Yang, Dilara Gokay, Skanda Koppula, Ankush Gupta, Joseph Heyward, Ignacio Rocco, Ross Goroshin, Jo ˜ao Carreira, et al. Bootstap: Boot- strapped training for tracking-any-point. arXiv preprint arXiv:2402.00847, 2024. 1, 2, 3

    2024 arXiv

  7. [15]

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

    Alexey Dosovitskiy. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 1, 8

    2010 arXiv

  8. [16]

    Kubric: A scalable dataset generator

    Klaus Greff, Francois Belletti, Lucas Beyer, Carl Doersch, Yilun Du, Daniel Duckworth, David J Fleet, Dan Gnanapra- gasam, Florian Golemo, Charles Herrmann, et al. Kubric: A scalable dataset generator. In Proceedings of the IEEE/CVF conference on computer vision and pattern re...

    2022

  9. [17]

    Asic: Aligning sparse in-the-wild image collections

    Kamal Gupta, Varun Jampani, Carlos Esteves, Abhinav Shri- vastava, Ameesh Makadia, Noah Snavely, and Abhishek Kar. Asic: Aligning sparse in-the-wild image collections. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4134–4145, 2023. 3

    2023

  10. [18]

    Unsupervised semantic segmentation by distilling feature correspondences

    Mark Hamilton, Zhoutong Zhang, Bharath Hariharan, Noah Snavely, and William T Freeman. Unsupervised semantic segmentation by distilling feature correspondences. arXiv preprint arXiv:2203.08414, 2022. 3

    2022 arXiv

  11. [19]

    Harley, Zhaoyuan Fang, and Katerina Fragkiadaki

    Adam W. Harley, Zhaoyuan Fang, and Katerina Fragkiadaki. Particle video revisited: Tracking through occlusions using point trajectories. In ECCV, 2022. 1, 2, 7

    2022

  12. [20]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 1, 4, 5, 6, 7

    2016

  13. [21]

    Mask r-cnn

    Kaiming He, Georgia Gkioxari, Piotr Doll ´ar, and Ross Gir- shick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017. 1

    2017

  14. [22]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll´ar, and Ross Girshick. Masked autoencoders are scalable vision learners. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 16000– 16009, 2022. 3

    2022

  15. [23]

    Inve: Interactive neural video editing

    Jiahui Huang, Leonid Sigal, Kwang Moo Yi, Oliver Wang, and Joon-Young Lee. Inve: Interactive neural video editing. arXiv preprint arXiv:2307.07663, 2023. 1

    2023 arXiv

  16. [24]

    Robust estimation of a location parameter

    Peter J Huber. Robust estimation of a location parameter. In Breakthroughs in statistics: Methodology and distribution , pages 492–518. Springer, 1992. 5

    1992

  17. [25]

    Co- tracker: It is better to track together

    Nikita Karaev, Ignacio Rocco, Benjamin Graham, Natalia Neverova, Andrea Vedaldi, and Christian Rupprecht. Co- tracker: It is better to track together. arXiv preprint arXiv:2307.07635, 2023. 1, 2, 3, 4, 5, 6, 7

    2023 arXiv

  18. [26]

    Co- tracker3: Simpler and better point tracking by pseudo- labelling real videos

    Nikita Karaev, Iurii Makarov, Jianyuan Wang, Natalia Neverova, Andrea Vedaldi, and Christian Rupprecht. Co- tracker3: Simpler and better point tracking by pseudo- labelling real videos. arXiv preprint arXiv:2410.11831 ,

    arXiv

  19. [27]

    The kinetics hu- man action video dataset

    Will Kay, Joao Carreira, Karen Simonyan, Brian Zhang, Chloe Hillier, Sudheendra Vijayanarasimhan, Fabio Viola, Tim Green, Trevor Back, Paul Natsev, et al. The kinetics hu- man action video dataset. arXiv preprint arXiv:1705.06950,

    arXiv

  20. [28]

    Beyond pick-and-place: Tackling robotic stacking of diverse shapes

    Alex X Lee, Coline Manon Devin, Yuxiang Zhou, Thomas Lampe, Konstantinos Bousmalis, Jost Tobias Springenberg, Arunkumar Byravan, Abbas Abdolmaleki, Nimrod Gileadi, David Khosid, et al. Beyond pick-and-place: Tackling robotic stacking of diverse shapes. In 5th Annual Confer- en...

    2021

  21. [29]

    Sfnet: Learning object-aware semantic correspon- dence

    Junghyup Lee, Dohyung Kim, Jean Ponce, and Bumsub Ham. Sfnet: Learning object-aware semantic correspon- dence. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 2278–2287,

  22. [30]

    Ground- ing image matching in 3d with mast3r

    Vincent Leroy, Yohann Cabon, and J´erˆome Revaud. Ground- ing image matching in 3d with mast3r. arXiv preprint arXiv:2406.09756, 2024. 2, 3, 4

    2024 arXiv

  23. [31]

    Taptr: Tracking any point with transformers as detection

    Hongyang Li, Hao Zhang, Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, and Lei Zhang. Taptr: Tracking any point with transformers as detection. arXiv preprint arXiv:2403.13042, 2024. 2, 3

    2024 arXiv

  24. [32]

    Tsm: Temporal shift module for efficient video understanding

    Ji Lin, Chuang Gan, and Song Han. Tsm: Temporal shift module for efficient video understanding. In Proceedings of the IEEE/CVF international conference on computer vision, pages 7083–7093, 2019. 2, 3, 5, 6, 7

    2019

  25. [33]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 1

    2021

  26. [34]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 5

    2017 arXiv

  27. [35]

    Im- proving semantic correspondence with viewpoint-guided spherical maps

    Octave Mariotti, Oisin Mac Aodha, and Hakan Bilen. Im- proving semantic correspondence with viewpoint-guided spherical maps. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19521– 19530, 2024. 3

    2024

  28. [36]

    Deep spectral methods: A surprisingly strong baseline for unsupervised semantic segmentation and localization

    Luke Melas-Kyriazi, Christian Rupprecht, Iro Laina, and Andrea Vedaldi. Deep spectral methods: A surprisingly strong baseline for unsupervised semantic segmentation and localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 8...

    2022

  29. [37]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023. 1, 2, ...

    2023 arXiv

  30. [38]

    Pytorch: An im- perative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An im- perative style, high-performance deep learning library. Ad- vances in neural information processing systems ...

    2019

  31. [39]

    The 2017 davis challenge on video object segmentation.arXiv preprint arXiv:1704.00675, 2017

    Jordi Pont-Tuset, Federico Perazzi, Sergi Caelles, Pablo Ar- bel´aez, Alex Sorkine-Hornung, and Luc Van Gool. The 2017 davis challenge on video object segmentation.arXiv preprint arXiv:1704.00675, 2017. 5

    2017 arXiv

  32. [40]

    Do vision trans- formers see like convolutional neural networks? Advances in neural information processing systems, 34:12116–12128,

    Maithra Raghu, Thomas Unterthiner, Simon Kornblith, Chiyuan Zhang, and Alexey Dosovitskiy. Do vision trans- formers see like convolutional neural networks? Advances in neural information processing systems, 34:12116–12128,

  33. [41]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 3

    2022

  34. [42]

    Efficient content-based sparse attention with rout- ing transformers

    Aurko Roy, Mohammad Saffar, Ashish Vaswani, and David Grangier. Efficient content-based sparse attention with rout- ing transformers. Transactions of the Association for Com- putational Linguistics, 9:53–68, 2021. 4

    2021

  35. [43]

    Time does tell: Self-supervised time- tuning of dense image representations

    Mohammadreza Salehi, Efstratios Gavves, Cees GM Snoek, and Yuki M Asano. Time does tell: Self-supervised time- tuning of dense image representations. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 16536–16547, 2023. 3

    2023

  36. [44]

    Learning universal semantic correspondences with no supervision and automatic data curation

    Aleksandar Shtedritski, Andrea Vedaldi, and Christian Rup- precht. Learning universal semantic correspondences with no supervision and automatic data curation. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 933–943, 2023. 3

    2023

  37. [45]

    Raft: Recurrent all-pairs field transforms for optical flow

    Zachary Teed and Jia Deng. Raft: Recurrent all-pairs field transforms for optical flow. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23– 28, 2020, Proceedings, Part II 16, pages 402–419. Springer,

    2020

  38. [46]

    Dino-tracker: Taming dino for self-supervised point track- ing in a single video

    Narek Tumanyan, Assaf Singer, Shai Bagon, and Tali Dekel. Dino-tracker: Taming dino for self-supervised point track- ing in a single video. In European Conference on Computer Vision, pages 367–385. Springer, 2025. 1, 3

    2025

  39. [47]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017. 4

    2017

  40. [48]

    Robotap: Tracking arbitrary points for few-shot visual imitation

    Mel Vecerik, Carl Doersch, Yi Yang, Todor Davchev, Yusuf Aytar, Guangyao Zhou, Raia Hadsell, Lourdes Agapito, and Jon Scholz. Robotap: Tracking arbitrary points for few-shot visual imitation. arXiv preprint arXiv:2308.15975, 2023. 1, 2

    2023 arXiv

  41. [49]

    Tracking everything everywhere all at once

    Qianqian Wang, Yen-Yu Chang, Ruojin Cai, Zhengqi Li, Bharath Hariharan, Aleksander Holynski, and Noah Snavely. Tracking everything everywhere all at once. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 19795–19806, 2023. 1, 3

    2023

  42. [50]

    Shape of motion: 4d reconstruc- tion from a single video

    Qianqian Wang, Vickie Ye, Hang Gao, Jake Austin, Zhengqi Li, and Angjoo Kanazawa. Shape of motion: 4d reconstruc- tion from a single video. arXiv preprint arXiv:2407.13764,

  43. [51]

    Rethinking self-supervised correspondence learning: A video frame-level similarity per- spective

    Jiarui Xu and Xiaolong Wang. Rethinking self-supervised correspondence learning: A video frame-level similarity per- spective. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 10075–10085, 2021. 3

    2021

  44. [52]

    Deconvolutional networks

    Matthew D Zeiler, Dilip Krishnan, Graham W Taylor, and Rob Fergus. Deconvolutional networks. In 2010 IEEE Com- puter Society Conference on computer vision and pattern recognition, pages 2528–2535. IEEE, 2010. 7

    2010

  45. [53]

    Scaling vision transformers

    Xiaohua Zhai, Alexander Kolesnikov, Neil Houlsby, and Lu- cas Beyer. Scaling vision transformers. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12104–12113, 2022. 1

    2022

  46. [54]

    Pointodyssey: A large-scale synthetic dataset for long-term point tracking

    Yang Zheng, Adam W Harley, Bokui Shen, Gordon Wet- zstein, and Leonidas J Guibas. Pointodyssey: A large-scale synthetic dataset for long-term point tracking. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision, pages 19855–19865, 2023. 1

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.