Pith. sign in

REVIEW 6 major objections 6 minor 61 references

Video Deblurring by Sharpness Prior Detection and Edge Information

T0 review · 6 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read This paper claims that making the sharp-frame frequency in a video deblurring dataset a tunable parameter, combined with a network that detects and transfers sharp-frame features and edge emphasis, improves average deblurring quality by…

desk verdict Useful variable-ratio sharp-frame dataset and a fast logistic-regression detector, but the headline SOTA claim is undercut by unfair baselines and an unreproducible Table 1. read the letter →

arxiv 2501.12246 v1 pith:4QD7PYZO submitted 2025-01-21 cs.CV

classification cs.CV
keywords videodeblurringsharpframedetectionGoProRSdatasetattention-basedencoder-decoderRichardson-Lucyedgeenhancementlogisticregressionfocusmetricsmotionblurdomainadaptation
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 argues that fixed sharp-to-blur ratios in video deblurring datasets hinder generalization, and that letting the ratio vary makes deblurring models more adaptable. To test this, the authors build GoProRS from GoPro footage by sampling window lengths so a user-chosen fraction $r$ of frames come out sharp, and train SPEINet, a network that detects nearby sharp frames with logistic regression on six classical autofocus metrics, emphasizes edges via Richardson-Lucy, and transfers sharp-frame features into an attention-based encoder-decoder. The paper reports that SPEINet trained on GoProRS reaches an average PSNR of 30.557 across GoProO, GoProS, GoProRS, and BSD, roughly 3.2% higher than recent baselines, and that its advantage grows as sharp-frame frequency increases. If correct, the work would mean variable sharp-frame availability is a key axis for video deblurring performance, not a nuisance to be fixed at a single ratio.

What carries the argument

The load-bearing machinery is the conditional transfer of sharp-frame features through a search-and-transfer module patterned on texture transformers. Sharp frames are located by a logistic-regression detector over six classical focus measures (MIS3, GRA7, LAP1, STA3, DCT3, WAV1); each detected sharp frame is encoded at three scales, while the blurry triplet and its Richardson-Lucy edge-emphasized version are encoded separately. A SwinIR block fuses the blur and edge encodings, a similarity search finds corresponding sharp-feature patches, and soft-attention refinement with cross-scale feature integration lets the decoder reconstruct the central frame. When no sharp frame is found, a self-search branch substitutes transfer with within-blur similarity.

What would settle it

Retrain SPEINet on the same data as each baseline, such as GoProO only, and evaluate on the same four datasets; if the average PSNR advantage over VRT or D2Net disappears or reverses, the central claim that the architecture plus GoProRS generalizes better is falsified. A second check is to measure detection accuracy of the logistic-regression detector on videos whose sharp-frame ratio falls outside the training distribution, such as $r=0.02$; if sharp detection fails there, the claimed robustness to variable ratios does not hold.

Watch

Extended reading notes

Core claim

The central claim is that a video deblurring model can be made both more accurate and more general by treating sharp frames as a tunable prior rather than a fixed assumption. Concretely, SPEINet detects the closest sharp frame among the previous seven frames using a logistic-regression classifier over six autofocus metrics, sharpens edges with Richardson-Lucy deconvolution, encodes blurry and sharp frames with local-attention residual blocks, and then transfers the sharp frame's features into the decoder through search-and-transfer and self-search branches. Trained on GoProRS with sharp ratios 0.1, 0.3, and 0.5, the model outperforms the compared state of the art on average and by a wider margin as the sharp ratio rises, which the authors interpret as evidence that the fixed 50% sharp ratio in GoProS biases models and hurts domain adaptation.

Load-bearing premise

The main load-bearing assumption is that averaging numbers across datasets with different training distributions is a fair comparison: most baselines were trained on GoProO or GoProS while SPEINet was trained on GoProRS, so the reported PSNR advantage is assumed to reflect the model rather than the training data.

Editorial extensions

If this is right

  • If SPEINet's claim is right, video deblurring models trained on variable sharp ratios generalize to both heavily blurred and real-world footage better than models trained on fixed ratios.
  • The logistic-regression sharp detector is fast enough to replace a learned Bi-LSTM detector: about 5.71 seconds versus 46 seconds per GoProRS sequence, with comparable final PSNR.
  • Edge emphasis from Richardson-Lucy contributes materially, with an ablation gain of roughly 4.36% PSNR over the baseline without sharpness or edge modules.
  • Performance scales with sharp-frame frequency: SPEINet's PSNR on GoProRS rises from 32.032 at $r=0.02$ to 35.452 at $r=0.5$, while most compared baselines stay flat or degrade.
  • A variable-ratio dataset such as GoProRS can be used to retrain existing sharp-frame-based models, improving their generalization across datasets outside their original training distribution.

Reading between the lines

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

  • The paper does not test what happens when all methods are trained on GoProRS; a re-run with matched training data would separate the dataset's contribution from the architecture's contribution.
  • A natural extension the paper leaves open is whether the detector's seven-frame search window is optimal for clustered bursts of sharp frames rather than randomly scattered ones.
  • The reported gains are averaged across datasets with different blur severity; per-dataset error bars would show whether the claimed advantage is consistent or driven by one test split.
  • The detector's accuracy advantage is partly inherited from training on the same distribution, so evaluating on videos whose sharp ratio falls outside the training set, such as $r=0.02$, would test the claimed robustness more strictly.
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

6 major / 6 minor

Summary. The paper introduces GoProRS, a new video deblurring dataset in which the proportion of sharp frames in a sequence is user-specified, and proposes SPEINet, a network that integrates a lightweight logistic-regression sharp-frame detector, a Richardson-Lucy edge-emphasis stage, and an attention-based encoder-decoder with search-transfer modules. The authors report that SPEINet, trained on GoProRS, achieves an average PSNR of 30.557 on GoProO, GoProS, GoProRS, and BSD, corresponding to a +3.2% average improvement over recent techniques in the abstract (the conclusion states +2.74%). The paper also includes ablations on the detection module and sub-module contributions.

Significance. If the claims are supported, the GoProRS dataset is a useful generalization of existing fixed-ratio sharp-frame datasets, and the lightweight classical-metric-based sharp detector is a practical contribution that could make sharp-frame prior methods more deployable. The paper is an empirical benchmark paper rather than a theoretical derivation, so the circularity burden is low; however, the central empirical claim currently rests on a comparison that is not controlled for training distribution, and the reported average numbers are not reproducible from the paper's own tables. These issues are fixable in principle, so the contribution is potentially significant but requires re-verification.

major comments (6)
  1. [§5.3, Table 1] The Average row in Table 1 cannot be reproduced from the per-dataset PSNR entries. For example, TSPNL's six PSNR values (32.9042, 32.611, 32.608, 26.573, 25.401, 26.522) have a simple mean of about 29.44, but the table reports 29.6033; VRT's values have a simple mean of about 29.03, but the table reports 28.5287. The paper does not state the averaging formula (e.g., frame-count-weighted or per-sequence averaging), and the discrepancy undermines the headline average improvement. Please provide the exact aggregation procedure and recompute the table, or correct the entries.
  2. [§5.3, Table 1] The central performance claim conflates architecture with training distribution. SPEINet is trained on GoProRS, while VRT, VDTR, TSP, TSPNL, DSTNet, and PVDNet are trained on GoProO or GoProS. On GoProO, SPEINet (31.527 dB) is well below VRT (34.79 dB), and on GoProS, the GoProS-trained D2Net(S) (35.554 dB) outperforms SPEINet (33.861 dB). The average advantage is therefore driven largely by the GoProRS and BSD columns, which favor the training distribution of SPEINet. To support the claim that the architecture is responsible for the improvement, at least one strong baseline (e.g., VRT or TSPNL) should be retrained on GoProRS under identical settings, or the comparison should be explicitly framed as a cross-dataset generalization study with the appropriate caveats.
  3. [§5.4, Table 2] The VDTR results are implausible and indicate an evaluation artifact: across all tested sharp-frame ratios, VDTR reports PSNR between 5.58 and 6.64 dB with SSIM between 0.0060 and 0.0105, which is worse than a constant-gray prediction and physically meaningless for a video deblurring transformer. This also contaminates the GoProRS row of Table 1, where VDTR is listed at 5.96 dB. Please rerun the VDTR evaluation with the correct model weights, data preprocessing, and metric computation, and update all affected tables and averages.
  4. [Abstract and §8 Conclusions] The abstract claims an average '+3.2% PSNR improvement', while the conclusions state 'an improvement of over 2.74%'. Both refer to the same set of experiments and the same average PSNR of 30.557, so the two numbers are mutually inconsistent. Moreover, the percentage is not defined: it is unclear whether it is relative to the best baseline, the average of all baselines, or a specific method such as D2Net(RS). Please state the reference baseline and recompute the percentage from the corrected Table 1.
  5. [§6.1, Tables 3 and 4] The comparison between the proposed logistic-regression detector (LD) and the Bi-LSTM detector is not matched on training data: the Bi-LSTM was trained on GoProS only, while LD is trained on GoProRS (Table 3) or GoProS (Table 4). The paper acknowledges this confound, but the sentence 'LD outperforms Bi-LSTM, improving detection accuracy by 15.56%' is still misleading, since the comparison in Table 3 is cross-dataset. Additionally, the 15.56% is a difference in percentage points (75.95 vs 60.39), not a 15.56% relative improvement. Please present a same-training-set comparison or clearly label the numbers as percentage-point differences.
  6. [§5.3, Table 1 and §6.2, Table 5] The ablation results in Table 5 report relative gains of 4.262% and 4.361% for the sharp-frame and edge modules, respectively. These percentages are not defined against a clear baseline: the rows compare different combinations of modules, but the baseline row (✗ ✗ ✗) has PSNR 29.592, and the +4.361% row (✗ ✓ ✓) has PSNR 33.953, which is a 14.7% relative increase, not 4.361%. Please clarify whether the percentages are relative improvements, absolute percentage-point increases, or something else, and make the computation consistent.
minor comments (6)
  1. [Abstract] The abstract contains typos: 'essential task' should be 'an essential task', and 'the the frequency' has a duplicated article.
  2. [§3, Eq. (1)] The indexing in Eq. (1) is confusing: the definition of the starting index σ(i)j is not self-contained, and the superscript on v is unclear. Please rewrite with explicit ranges and clarify how windows are non-overlapping.
  3. [§5.4, Table 2 caption] The caption says 'the other method were trained on GoProO'; the grammar should be 'the other methods were trained on GoProO'.
  4. [§8 Conclusions] The phrase 'an logistic-regression-based detector' should be 'a logistic-regression-based detector'.
  5. [§2 Related works] Several references are cited without complete context (e.g., [12], [13], [41]). Please ensure all citations have full bibliographic entries and that the described contributions match the cited works.
  6. [§7 Visualization] Figure 7 seems to show only six frames, but the text says 'six blur frames' and the figure caption says 'six frames'—please clarify whether these are frames from one sequence or separate sequences, and label the rows/columns accordingly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's contributions are a new dataset, a supervised sharp-frame detector, and an empirical benchmark; no claimed result is defined as, or fitted to, its own output.

full rationale

This is an empirical computer-vision benchmark paper rather than a derivation with fitted constants, so the circularity burden is low and no load-bearing circular step is found. The GoProRS dataset is constructed by an explicit generative rule (Eq. 1): window lengths are sampled, sharp labels are defined by l_j = (w_j <= 5), and ground-truth frames are the central frames of the averaged windows. The sharp-frame detector is a logistic regression over six classical autofocus metrics, trained in a supervised manner on ground-truth sharp/blur labels and evaluated on held-out test sequences; its accuracy (Table 3: LD 75.95% vs Bi-LSTM 60.39%) is an independent measurement, not a fitted replay of the deblurring metric. The deblurring model SPEINet is trained on GoProRS and evaluated on GoProO, GoProS, GoProRS, and BSD, with no target PSNR used as a training signal. The main performance claim, an average +3.2% PSNR improvement, rests on cross-dataset comparison: SPEINet is trained on GoProRS while most baselines were trained on GoProO or GoProS. That is a legitimate experimental-design and fairness concern, and Table 1's average row also appears numerically inconsistent with simple means of the listed entries, but neither issue is a self-definitional or fitted-input circularity. There are no load-bearing self-citations: references such as D2Net [38] and the texture-transformer search-transfer block [51] are external prior work whose components are adopted and ablated, not invoked as unverified uniqueness theorems. Hence the paper's derivation chain is self-contained with respect to circularity, and the appropriate score is 0.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central empirical claims rest on standard supervised learning assumptions and several hand-set hyperparameters (r, the sharp-label window threshold, lookback gamma, and metric kernel size). No new physical entities or speculative constructs are introduced.

free parameters (4)
  • sharp-frame ratio r = training: 0.1, 0.3, 0.5; evaluation: 0.02 to 0.5
    User-defined dataset parameter controlling the probability that a sampled window length is at most 5. The training ratios are chosen by hand rather than derived from first principles.
  • sharp/blur window threshold = w <= 5
    Frames averaged from windows of length at most 5 are labeled sharp, which is an arbitrary binary cutoff in Eq. (1). This definition directly sets the labels used to train and evaluate the detector.
  • lookback window gamma = 7
    The sharp detector only considers the previous 7 frames (Section 4.1.1). The text says larger values are not advisable but provides no quantitative analysis.
  • autofocus metric kernel size = 11
    Kernel size for the six autofocus metrics is set to 11 in Section 5.2; Appendix E tests k=3,5,7,11,51 and 11 is selected for the main experiments.
assumptions (3)
  • domain assumption Averaging high-frame-rate video frames produces realistic motion blur, following the GoPro protocol.
    Used to construct GoProRS from GoProO in Section 3; assumes synthetic averaging transfers to real-world blur sufficiently well.
  • domain assumption Classical autofocus measures plus logistic regression can separate sharp from blurred frames accurately enough for deblurring.
    Stage 1 (Section 4.1.1) relies entirely on six scalar focus metrics and a trained linear classifier; if this fails on a test distribution, the sharp prior information is unavailable.
  • domain assumption Richardson-Lucy deconvolution with a uniform blur assumption improves edge extraction on non-uniform video blur.
    The edge emphasis stage (Section 4.1.2) applies RL to blurred frames, although RL assumes a known uniform point spread function and can amplify artifacts on non-uniform dynamic blur.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Video Deblurring by Sharpness Prior Detection and Edge Information." pith.science (2026). https://pith.science/paper/4QD7PYZO

@misc{pith2026250112246,
  author       = {Pith},
  title        = {Pith review of: Video Deblurring by Sharpness Prior Detection and Edge Information},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4QD7PYZO}},
  note         = {Machine review of arXiv:2501.12246}
}
read the original abstract

Video deblurring is essential task for autonomous driving, facial recognition, and security surveillance. Traditional methods directly estimate motion blur kernels, often introducing artifacts and leading to poor results. Recent approaches utilize the detection of sharp frames within video sequences to enhance deblurring. However, existing datasets rely on fixed number of sharp frames, which may be too restrictive for some applications and may introduce a bias during model training. To address these limitations and enhance domain adaptability, this work first introduces GoPro Random Sharp (GoProRS), a new dataset where the the frequency of sharp frames within the sequence is customizable, allowing more diverse training and testing scenarios. Furthermore, it presents a novel video deblurring model, called SPEINet, that integrates sharp frame features into blurry frame reconstruction through an attention-based encoder-decoder architecture, a lightweight yet robust sharp frame detection and an edge extraction phase. Extensive experimental results demonstrate that SPEINet outperforms state-of-the-art methods across multiple datasets, achieving an average of +3.2% PSNR improvement over recent techniques. Given such promising results, we believe that both the proposed model and dataset pave the way for future advancements in video deblurring based on the detection of sharp frames.

Figures

Figures reproduced from arXiv: 2501.12246 by the authors.

Figure 1
Figure 1. Overview the SPEINet framework during the inference of the model. [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Schema of Sharp Frames Detection. Logistic regression is trained to perform binary [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Edge extraction obtained by applying the regularized-edge-emphasizing module [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Three blocks encoder module. The encoder receives the blur frames xi, the empha￾sized edges frames ci, and any closest sharp frame si. Here εθ(·, k) denotes the encoder until block k. The triplets ui and vi are deduced by applying ε(·, 3), frame-wise, to xi and ci resp…
Figure 5
Figure 5. Figure 5: Detail of Search-Transfer [51] and Self-Search modules. Bis is based on the maximum [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Details of attention-based decoder. Soft-Att refers to the Soft-Attention mechanisms [51] and CSFI refers to cross-scale feature integration module [51]. Com￾bined with up-sampling modules, continually capture contextual detail information to aid in the reconstruction …
Figure 7
Figure 7. Figure 7: Sample outputs produced by SPEINet and other state-of-the-art models on six frames taken from GoProRS (r=0.02). Blur SPEINet CDVD-TSP D2Nets DTSNet Ground truth 1ms-8ms 2ms-16ms 3ms-24ms [PITH_FULL_IMAGE:figures/full_fig_p019_7.png]
Figure 8
Figure 8. Figure 8: Examples of outputs produced by SPEINet and other state-of-the-art models on three samples sequences taken from the BSD real-world dataset. 8. Conclusions This work presented a novel video deblurring model called Sharpness Prior Detector and Edge Information Network (S…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 59 canonical work pages

  1. [1]

    Medical image enhancement and deblurring, in: Proceedings of the 11th National Techni- cal Seminar on Unmanned System Technology 2019: NUSYS’19, Springer Singapore

    Amini Gougeh, R., Yousefi Rezaii, T., Farzamnia, A., 2021. Medical image enhancement and deblurring, in: Proceedings of the 11th National Techni- cal Seminar on Unmanned System Technology 2019: NUSYS’19, Springer Singapore. pp. 543–554

  2. [2]

    Vdtr: Video deblurring with transformer

    Cao, M., Fan, Y., Zhang, Y., Wang, J., Yang, Y., 2022. Vdtr: Video deblurring with transformer. IEEE Transactions on Circuits and Systems for Video Technology 33, 160–171

  3. [3]

    Chakravarthi, B., Verma, A.A., Daniilidis, K., Fermuller, C., Yang, Y.,

  4. [4]

    Hierarchical integration diffusion model for realistic image deblurring, in: Advances in Neural Information Processing Systems

    Chen, Z., Zhang, Y., Liu, D., Gu, J., Kong, L., Yuan, X., 2024. Hierarchical integration diffusion model for realistic image deblurring, in: Advances in Neural Information Processing Systems

  5. [5]

    Video deblurring and flow- guided feature aggregation for obstacle detection in agricultural videos

    Cheng, K., Zhu, X., Zhan, Y., Pei, Y., 2022. Video deblurring and flow- guided feature aggregation for obstacle detection in agricultural videos. International Journal of Multimedia Information Retrieval 11, 577–588. 20

  6. [6]

    Dual-domain attention for image deblurring, in: Proceedings of the AAAI Conference on Artificial Intelligence, pp

    Cui, Y., Tao, Y., Ren, W., Knoll, A., 2023. Dual-domain attention for image deblurring, in: Proceedings of the AAAI Conference on Artificial Intelligence, pp. 479–487

  7. [7]

    Hand-held video deblurring via efficient fourier aggregation

    Delbracio, M., Sapiro, G., 2015. Hand-held video deblurring via efficient fourier aggregation. IEEE Transactions on Computational Imaging 1, 270– 283

  8. [8]

    Multi-scale residual low- pass filter network for image deblurring, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Dong, J., Pan, J., Yang, Z., Tang, J., 2023. Multi-scale residual low- pass filter network for image deblurring, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 12345–12354

Show all 61 references
  1. [9]

    Domain-adaptive video deblurring via test-time blurring, in: European Conference on Computer Vision, Springer

    He, J.T., Tsai, F.J., Wu, J.H., Peng, Y.T., Tsai, C.C., Lin, C.W., Lin, Y.Y., 2025. Domain-adaptive video deblurring via test-time blurring, in: European Conference on Computer Vision, Springer. pp. 125–142

  2. [10]

    Effective video deblurring based on feature-enhanced deep learning network for daytime and nighttime images

    Huang, D.Y., Chen, C.H., Chen, T.Y., Li, J.E., Hsiao, H.L., Wang, D.J., Wen, C.K., 2024. Effective video deblurring based on feature-enhanced deep learning network for daytime and nighttime images. Multimedia Tools and Applications , 1–27

  3. [11]

    Generalized video deblurring for dynamic scenes, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Hyun Kim, T., Mu Lee, K., 2015. Generalized video deblurring for dynamic scenes, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5426–5434

  4. [12]

    Stereoscopic video deblurring transformer

    Imani, H., Islam, M.B., Junayed, M.S., Ahad, M.A.R., 2024. Stereoscopic video deblurring transformer. Scientific Reports 14, 14342

  5. [13]

    Frequency-aware event-based video deblurring for real-world motion blur, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Kim, T., Cho, H., Yoon, K.J., 2024. Frequency-aware event-based video deblurring for real-world motion blur, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 24966–24976

  6. [14]

    Event-guided deblurring of unknown exposure time videos, in: European Conference on Computer Vision, Springer Nature Switzerland, Cham

    Kim, T., Lee, J., Wang, L., Yoon, K.J., 2022. Event-guided deblurring of unknown exposure time videos, in: European Conference on Computer Vision, Springer Nature Switzerland, Cham. pp. 519–538

  7. [15]

    Efficient frequency domain-based transformers for high-quality image deblurring, in: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, pp

    Kong, L., Dong, J., Ge, J., Li, M., Pan, J., 2023. Efficient frequency domain-based transformers for high-quality image deblurring, in: Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, pp. 5886–5895

  8. [16]

    Vrt: A video restoration transformer

    Liang, J., Cao, J., Fan, Y., Zhang, K., Ranjan, R., Li, Y., Timofte, R., Van Gool, L., 2024. Vrt: A video restoration transformer. IEEE Transac- tions on Image Processing

  9. [17]

    Swinir: Image restoration using swin transformer, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Liang, J., Cao, J., Sun, G., Zhang, K., Van Gool, L., Timofte, R., 2021. Swinir: Image restoration using swin transformer, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 1833–1844. 21

  10. [18]

    Lightvid: Efficient video deblurring with spatial-temporal feature fusion

    Lin, L., Wei, G., Liu, K., Feng, W., Zhao, T., 2024. Lightvid: Efficient video deblurring with spatial-temporal feature fusion. IEEE Transactions on Circuits and Systems for Video Technology

  11. [19]

    Lin, S., Zhang, J., Pan, J., Jiang, Z., Zou, D., Wang, Y., Ren, J., 2020. Learning event-driven video deblurring and interpolation, in: Computer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 2328, 2020, Proceedings, Part VIII, Springer International Publish...

  12. [20]

    Intriguing findings of frequency selection for image deblurring, in: Proceedings of the AAAI Conference on Artificial Intelligence, pp

    Mao, X., Liu, Y., Liu, F., Li, Q., Shen, W., Wang, Y., 2023. Intriguing findings of frequency selection for image deblurring, in: Proceedings of the AAAI Conference on Artificial Intelligence, pp. 1905–1913

  13. [21]

    Theory of edge detection

    Marr, D., Hildreth, E., 1980. Theory of edge detection. Proceedings of the Royal Society of London. Series B. Biological Sciences 207, 187–217

  14. [22]

    Nah, S., Baik, S., Hong, S., Moon, G., Son, S., Timofte, R., Mu Lee, K., 2019. Ntire 2019 challenge on video deblurring and super-resolution: Dataset and study, in: Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition Workshops, pp. 0–0

  15. [23]

    Deep multi-scale convolutional neural network for dynamic scene deblurring, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Nah, S., Hyun Kim, T., Mu Lee, K., 2017. Deep multi-scale convolutional neural network for dynamic scene deblurring, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3883–3891

  16. [24]

    A comprehensive review on deep learning-based methods for video anomaly detection

    Nayak, R., Pati, U.C., Das, S.K., 2021. A comprehensive review on deep learning-based methods for video anomaly detection. Image and Vision Computing 106, 104078

  17. [25]

    Cascaded deep video deblurring using temporal sharpness prior, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Pan, J., Bai, H., Tang, J., 2020. Cascaded deep video deblurring using temporal sharpness prior, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3043–3051

  18. [26]

    Cascaded deep video deblurring using temporal sharpness prior and non-local spatial-temporal similarity

    Pan, J., Xu, B., Bai, H., Tang, J., Yang, M.H., 2023a. Cascaded deep video deblurring using temporal sharpness prior and non-local spatial-temporal similarity. IEEE Transactions on Pattern Analysis and Machine Intelligence 45, 9411–9425

  19. [27]

    Deep discriminative spa- tial and temporal network for efficient video deblurring, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Pan, J., Xu, B., Dong, J., Ge, J., Tang, J., 2023b. Deep discriminative spa- tial and temporal network for efficient video deblurring, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 22191–22200

  20. [28]

    Simultaneous stereo video de- blurring and scene flow estimation, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Pan, L., Dai, Y., Liu, M., Porikli, F., 2017. Simultaneous stereo video de- blurring and scene flow estimation, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4382–4391

  21. [29]

    Analysis of focus measure opera- tors for shape-from-focus

    Pertuz, S., Puig, D., Garcia, M.A., 2013. Analysis of focus measure opera- tors for shape-from-focus. Pattern Recognition 46, 1415–1432. 22

  22. [30]

    Rethinking video deblurring with wavelet- aware dynamic transformer and diffusion model, in: European Conference on Computer Vision, Springer

    Rao, C., Li, G., Lan, Z., Sun, J., Luan, J., Xing, W., Zhao, L., Lin, H., Dong, J., Zhang, D., 2025. Rethinking video deblurring with wavelet- aware dynamic transformer and diffusion model, in: European Conference on Computer Vision, Springer. pp. 421–437

  23. [31]

    Mul- tiscale structure guided diffusion for image deblurring, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Ren, M., Delbracio, M., Talebi, H., Gerig, G., Milanfar, P., 2023. Mul- tiscale structure guided diffusion for image deblurring, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 10721– 10733

  24. [32]

    Fast ultra high-definition video deblurring via multi-scale separable network

    Ren, W., Deng, S., Zhang, K., Song, F., Cao, X., Yang, M.H., 2024. Fast ultra high-definition video deblurring via multi-scale separable network. International Journal of Computer Vision 132, 1817–1834

  25. [33]

    Video deblurring via semantic segmentation and pixel-wise non-linear kernel, in: Proceedings of the IEEE International Conference on Computer Vision, pp

    Ren, W., Pan, J., Cao, X., Yang, M.H., 2017. Video deblurring via semantic segmentation and pixel-wise non-linear kernel, in: Proceedings of the IEEE International Conference on Computer Vision, pp. 1077–1085

  26. [34]

    Bayesian-based iterative method of image restora- tion

    Richardson, W.H., 1972. Bayesian-based iterative method of image restora- tion. JoSA 62, 55–59

  27. [35]

    Video restoration based on deep learning: a comprehensive survey

    Rota, C., Buzzelli, M., Bianco, S., Schettini, R., 2023. Video restoration based on deep learning: a comprehensive survey. Artificial Intelligence Review 56, 5317–5364

  28. [36]

    Deep residual learning for image recognition: A survey

    Shafiq, M., Gu, Z., 2022. Deep residual learning for image recognition: A survey. Applied Sciences 12, 8972

  29. [37]

    Aggregating nearest sharp features via hybrid transformers for video deblurring

    Shang, W., Ren, D., Yang, Y., Zuo, W., 2025. Aggregating nearest sharp features via hybrid transformers for video deblurring. Information Sciences 694, 121689

  30. [38]

    Bringing events into video deblurring with non-consecutively blurry frames, in: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Shang, W., Ren, D., Zou, D., Ren, J.S., Luo, P., Zuo, W., 2021. Bringing events into video deblurring with non-consecutively blurry frames, in: Pro- ceedings of the IEEE/CVF International Conference on Computer Vision, pp. 4531–4540

  31. [39]

    Spatial temporal video enhancement using alternating exposures

    Shen, W., Cheng, M., Lu, G., Zhai, G., Chen, L., Asif, M.S., Gao, Z., 2021. Spatial temporal video enhancement using alternating exposures. IEEE Transactions on Circuits and Systems for Video Technology 32, 4912–4926

  32. [40]

    A 3x3 isotropic gradient operator for image processing

    Sobel, I., Feldman, G., et al., 1968. A 3x3 isotropic gradient operator for image processing. a talk at the Stanford Artificial Project in 1968, 271–272

  33. [41]

    Recurrent video deblurring with blur-invariant motion estimation and pixel volumes

    Son, H., Lee, J., Lee, J., Cho, S., Lee, S., 2021. Recurrent video deblurring with blur-invariant motion estimation and pixel volumes. ACM Transac- tions on Graphics (TOG) 40, 1–18

  34. [42]

    Memory-based gradient-guided progressive propagation network for video deblurring

    Song, G., Gai, S., Da, F., 2024. Memory-based gradient-guided progressive propagation network for video deblurring. The Visual Computer , 1–16. 23

  35. [43]

    Deep video deblurring for hand-held cameras, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

    Su, S., Delbracio, M., Wang, J., Sapiro, G., Heidrich, W., Wang, O., 2017. Deep video deblurring for hand-held cameras, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1279–1288

  36. [44]

    Gated spatio-temporal attention- guided video deblurring, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Suin, M., Rajagopalan, A.N., 2021. Gated spatio-temporal attention- guided video deblurring, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 7802–7811

  37. [45]

    High-resolution representations for labeling pixels and regions

    Sun, K., Zhao, Y., Jiang, B., Cheng, T., Xiao, B., Liu, D., Mu, Y., Wang, X., Liu, W., Wang, J., 2019. High-resolution representations for labeling pixels and regions. arXiv preprint arXiv:1904.04514

  38. [46]

    Long-term object tracking based on joint tracking and detection strategy with siamese net- work

    Sun, L., Zhang, J., Yang, Z., Gao, D., Fan, B., 2024. Long-term object tracking based on joint tracking and detection strategy with siamese net- work. Multimedia Systems 30, 162

  39. [47]

    Video deblurring via spatiotemporal pyramid network and adversarial gradient prior

    Wang, T., Zhang, X., Jiang, R., Zhao, L., Chen, H., Luo, W., 2021. Video deblurring via spatiotemporal pyramid network and adversarial gradient prior. Computer Vision and Image Understanding 203, 103135

  40. [48]

    Xian, R., Wang, X., Manocha, D., 2024. Mitfas: Mutual information based temporal feature alignment and sampling for aerial video action recognition, in: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pp. 6625–6634

  41. [49]

    Deep video deblurring using sharpness features from exemplars

    Xiang, X., Wei, H., Pan, J., 2020. Deep video deblurring using sharpness features from exemplars. IEEE Transactions on Image Processing 29, 8976– 8987

  42. [50]

    Sharp- former: Learning local feature preserving global representations for image deblurring

    Yan, Q., Gong, D., Wang, P., Zhang, Z., Zhang, Y., Shi, J.Q., 2023. Sharp- former: Learning local feature preserving global representations for image deblurring. IEEE Transactions on Image Processing 32, 2857–2866

  43. [51]

    Learning texture transformer network for image super-resolution, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Yang, F., Yang, H., Fu, J., Lu, H., Guo, B., 2020. Learning texture transformer network for image super-resolution, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5791–5800

  44. [52]

    Blur-aware spatio-temporal sparse trans- former for video deblurring, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Zhang, H., Xie, H., Yao, H., 2024. Blur-aware spatio-temporal sparse trans- former for video deblurring, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 2673–2681

  45. [53]

    Zhang, K., Ren, W., Luo, W., Lai, W.S., Stenger, B., Yang, M.H., Li, H.,

  46. [54]

    Zhong, Z., Gao, Y., Zheng, Y., Zheng, B., 2020. Efficient spatio-temporal recurrent neural network for video deblurring, in: Computer Vision – ECCV 2020: 16th European Conference, Glasgow, UK, August 2328, 2020, Proceedings, Part VI, Springer International Publishing. pp. 191–207. 24

  47. [55]

    Dast-net: Depth-aware spatio-temporal network for video deblurring, in: 2022 IEEE International Conference on Multimedia and Expo (ICME), IEEE

    Zhu, Q., Xiao, Z., Huang, J., Zhao, F., 2022. Dast-net: Depth-aware spatio-temporal network for video deblurring, in: 2022 IEEE International Conference on Multimedia and Expo (ICME), IEEE. pp. 1–6

  48. [56]

    Video Deblurring by Sharpness Prior Detection and Edge Information

    Zhu, Q., Zhou, M., Zheng, N., Li, C., Huang, J., Zhao, F., 2023. Exploring temporal frequency spectrum in deep video deblurring, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 12428– 12437. 25 Appendix of “Video Deblurring by Sharpness Prior D...

  49. [59]

    Nevertheless, the code is not publicly available

    The model proposed in [10] has not be tested on the GoProO dataset. Nevertheless, the code is not publicly available

  50. [60]

    The model proposed in [13] leverages extra information deduced from an event camera, and hence can not be directly compared to SPEINet

  51. [61]

    Appendix C

    The model proposed in [12] focused on Stereo Blur dataset, and the code is not publicly available. Appendix C. Video frames of GoProRS and GoProS In this section, we compared the number of each video sharp frames in GoProRS and GoProS. As shown in Table C.7, The number of vide...

  52. [2022]

    International Journal of Computer Vision 130, 2103–2130

    Deep image deblurring: A survey. International Journal of Computer Vision 130, 2103–2130

  53. [2024]

    arXiv preprint arXiv:2408.13627

    Recent event camera innovations: A survey. arXiv preprint arXiv:2408.13627

Pith tools

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