pith. sign in

arxiv: 2508.13401 · v3 · submitted 2025-08-18 · 💻 cs.CV

AIM 2025 Rip Current Segmentation (RipSeg) Challenge Report

Andrei Dumitriu , Florin Miron , Florin Tatui , Radu Tudor Ionescu , Radu Timofte , Aakash Ralhan , Florin-Alexandru Vasluianu , Shenyang Qian
show 19 more authors
Mitchell Harley Imran Razzak Yang Song Pu Luo Yumei Li Cong Xu Jinming Chai Kexin Zhang Licheng Jiao Lingling Li Siqi Yu Chao Zhang Kehuan Song Fang Liu Puhua Chen Xu Liu Jin Hu Jinyang Xu Biao Liu
This is my paper

Pith reviewed 2026-05-18 21:56 UTC · model grok-4.3

classification 💻 cs.CV
keywords rip current segmentationdeep learningdomain adaptationinstance segmentationRipVIS datasetbeach safetycomputer visionchallenge report
0
0 comments X

The pith

Top entries in the RipSeg challenge succeed by pairing deep learning with domain adaptation on the diverse RipVIS dataset.

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

This report summarizes the first AIM 2025 RipSeg Challenge, which tested automatic segmentation of rip currents in still images using the RipVIS dataset. The dataset includes varied locations, rip current types, and camera angles to create a realistic benchmark for a safety-critical task. Five valid submissions competed under a composite metric of F1, F2, AP50, and AP[50:95]. The strongest results came from deep learning models that incorporated domain adaptation, pretrained weights, and generalization methods to handle condition changes. The overview also notes remaining challenges and directions for future work on this underexplored problem.

Core claim

The top-performing methods leveraged deep learning architectures, domain adaptation techniques, pretrained models, and domain generalization strategies to improve performance under diverse conditions on the RipVIS dataset.

What carries the argument

The composite evaluation metric that combines F1, F2, AP50, and AP[50:95] to produce a single ranking from the five valid test submissions.

If this is right

  • Models trained this way can delineate rip current boundaries more precisely across different beaches and viewpoints.
  • Pretrained models reduce the data needs for this specialized segmentation task.
  • Domain adaptation allows a single system to maintain accuracy when camera setups or water conditions change.
  • The benchmark creates a starting point for adding temporal information from video sequences.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same adaptation strategies could transfer to segmenting other dynamic water features such as waves or foam lines.
  • Deploying the top models in fixed beach cameras would let safety teams receive alerts based on exact current outlines rather than rough location estimates.
  • Expanding the dataset with nighttime or stormy images would test whether the current adaptation techniques still hold.

Load-bearing premise

The five submissions and the chosen composite metric give a reliable picture of what works best for real beach monitoring.

What would settle it

A new method that ranks first on the same test set while using neither deep learning nor any domain adaptation or generalization step would undermine the reported pattern of success.

Figures

Figures reproduced from arXiv: 2508.13401 by Aakash Ralhan, Andrei Dumitriu, Biao Liu, Chao Zhang, Cong Xu, Fang Liu, Florin-Alexandru Vasluianu, Florin Miron, Florin Tatui, Imran Razzak, Jin Hu, Jinming Chai, Jinyang Xu, Kehuan Song, Kexin Zhang, Licheng Jiao, Lingling Li, Mitchell Harley, Puhua Chen, Pu Luo, Radu Timofte, Radu Tudor Ionescu, Shenyang Qian, Siqi Yu, Xu Liu, Yang Song, Yumei Li.

Figure 1
Figure 1. Figure 1: Examples from the RipVIS dataset [17], which also forms the basis of the RipSeg Challenge. The four columns illustrate different camera orientations: (a) aerial bird’s-eye, (b) aerial tilted, (c) elevated beachfront, and (d) water-level beachfront. The examples highlight the diversity of rip currents across locations, types, and viewpoints. Rip currents are visible through disrupted wave-breaking patterns,… view at source ↗
Figure 2
Figure 2. Figure 2: Methodology of RipEye, using HRDA [22] and SHADE [57]. without rip currents were named RipSeg-NR-<number>. For the final test set, all files have been renamed with a ran￾domized hash as an extra step in preventing fraud. After the two days, participants had to submit a description of their team, their reproducible code and a description of their ap￾proach. Only teams that passed this final check were con￾s… view at source ↗
Figure 3
Figure 3. Figure 3: Overview of team RipSense’s workflow showing the predicted segmentation mask of a random test sample, using SparseInst as [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Rip current segmentation approach by team Gogogochufalou. Raw data is augmented, then processed by SparseInst (outputting [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Team ZYS’s scheme for rip current segmentation, using YOLO11x. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Team Simplehh’s fine-tuned YOLOv8n Pipeline. Four CBAM blocks were inserted after the C2f block in the Neck part to [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
read the original abstract

This report presents an overview of the AIM 2025 RipSeg Challenge, a competition designed to advance techniques for automatic rip current segmentation in still images. Rip currents are dangerous, fast-moving flows that pose a major risk to beach safety worldwide, making accurate visual detection an important and underexplored research task. The challenge builds on RipVIS, the largest available rip current dataset, and focuses on single-class instance segmentation, where precise delineation is critical to fully capture the extent of rip currents. The dataset spans diverse locations, rip current types, and camera orientations, providing a realistic and challenging benchmark. In total, $75$ participants registered for this first edition, resulting in $5$ valid test submissions. Teams were evaluated on a composite score combining $F_1$, $F_2$, $AP_{50}$, and $AP_{[50:95]}$, ensuring robust and application-relevant rankings. The top-performing methods leveraged deep learning architectures, domain adaptation techniques, pretrained models, and domain generalization strategies to improve performance under diverse conditions. This report outlines the dataset details, competition framework, evaluation metrics, and final results, providing insights into the current state of rip current segmentation. We conclude with a discussion of key challenges, lessons learned from the submissions, and future directions for expanding RipSeg.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. This paper is a report on the AIM 2025 RipSeg Challenge for automatic rip current segmentation in images. It introduces the task, describes the RipVIS dataset with its diversity in locations, types, and camera orientations, outlines the challenge rules with 75 registrations and 5 valid submissions, specifies the composite metric of F1, F2, AP50 and AP[50:95], and summarizes that the leading approaches employed deep learning with domain adaptation and generalization techniques. The report also covers lessons learned and future directions.

Significance. This challenge report holds significance by creating a public benchmark for an important but underexplored computer vision task related to beach safety. The use of the largest rip current dataset and a multi-metric evaluation provides a solid foundation for comparing methods. If the performance claims and technique attributions are accurate, it can accelerate progress in domain-robust segmentation models.

major comments (1)
  1. [Abstract] The claim regarding the techniques used by top-performing methods (deep learning architectures, domain adaptation, pretrained models, and domain generalization strategies) requires more substantiation. With only five valid submissions and high-level summaries provided, there is no detailed breakdown of each method's components, no ablation studies, and no analysis of performance on specific challenging subsets of the RipVIS test set to confirm that these strategies drove the improvements under diverse conditions.
minor comments (1)
  1. The report could include a table listing the five submissions with their reported scores and key methodological highlights to improve readability and allow readers to better assess the results.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback on the AIM 2025 RipSeg Challenge report. We address the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] The claim regarding the techniques used by top-performing methods (deep learning architectures, domain adaptation, pretrained models, and domain generalization strategies) requires more substantiation. With only five valid submissions and high-level summaries provided, there is no detailed breakdown of each method's components, no ablation studies, and no analysis of performance on specific challenging subsets of the RipVIS test set to confirm that these strategies drove the improvements under diverse conditions.

    Authors: We agree that the abstract statement is high-level. It is grounded in the method descriptions submitted by the five participating teams, which are summarized in the main body of the report (Section on submissions). Each top team explicitly reported using deep learning backbones, domain adaptation or generalization modules, and pretrained weights. However, as this is a challenge report rather than a methods paper, we did not reproduce ablations or conduct new per-subset analyses on the test set. To address the concern, we will revise the abstract to qualify the claim as 'as reported by the participating teams' and add a concise table in the results section that tabulates the key components declared by each submission. We cannot add new ablation studies or subset-specific performance breakdowns without additional experiments outside the scope of the original challenge submissions. revision: partial

Circularity Check

0 steps flagged

No circularity: factual challenge report with no derivations or self-referential claims

full rationale

The manuscript is a competition summary report describing the RipSeg challenge setup, the RipVIS dataset, evaluation metrics (F1, F2, AP50, AP[50:95]), and high-level outcomes from five external submissions. No equations, predictions, fitted parameters, or derivation chains appear anywhere in the text. The statement that top methods 'leveraged deep learning architectures, domain adaptation techniques, pretrained models, and domain generalization strategies' is presented as a descriptive observation of participant entries rather than a derived result or self-citation load-bearing premise. Because the document contains no claimed first-principles results or reductions that collapse to its own inputs, the circularity score is 0 and the analysis is self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are introduced; the document is an administrative summary of a benchmark competition.

pith-pipeline@v0.9.0 · 5875 in / 955 out tokens · 31678 ms · 2026-05-18T21:56:35.648758+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

58 extracted references · 58 canonical work pages · 2 internal anchors

  1. [1]

    Arun Kumar and K.V .S.R

    S.V .V . Arun Kumar and K.V .S.R. Prasad. Rip current-related fatalities in India: a new predictive risk scale for forecasting rip currents. Natural Hazards, 70:313–335, 2014. 6

    work page 2014

  2. [2]

    Brander, Dale Dominey-Howes, C

    R. Brander, Dale Dominey-Howes, C. Champion, O. Del Vecchio, and B. Brighton. Brief Communication: A new perspective on the Australian rip current hazard. Nat- ural Hazards and Earth System Sciences, 13(6):1687–1690,

    work page

  3. [3]

    Brander and A.D

    Robert W. Brander and A.D. Short. Morphodynamics of a large-scale rip current system at Muriwai Beach, New Zealand. Marine Geology, 165(1-4):27–39, 2000. 2

    work page 2000

  4. [4]

    Chris Brewster, Richard E

    B. Chris Brewster, Richard E. Gould, and Robert W. Brander. Estimations of rip current rescues and drowning in the United States. Natural Hazards and Earth System Sciences , 19(2): 389–397, 2019. 2

    work page 2019

  5. [5]

    Castelle, Tim Scott, R.W

    B. Castelle, Tim Scott, R.W. Brander, and R.J. McCarroll. Rip current types, circulation and hazard. Earth-Science Re- views, 163:1–21, 2016. 2

    work page 2016

  6. [6]

    Sparse instance activation for real-time instance segmentation, 2022

    Tianheng Cheng, Xinggang Wang, Shaoyu Chen, Wenqiang Zhang, Qian Zhang, Chang Huang, Zhaoxiang Zhang, and Wenyu Liu. Sparse instance activation for real-time instance segmentation, 2022. 4, 5

    work page 2022

  7. [7]

    Explainable Rip Current Detection and Visualization with XAI EigenCAM

    Juno Choi, Muralidharan Rajendran, and Yong Cheol Suh. Explainable Rip Current Detection and Visualization with XAI EigenCAM. In Proceedings of 26th International Con- ference on Advanced Communications Technology, pages 1– 6, 2024. 2

    work page 2024

  8. [8]

    AIM 2025 high FPS non-uniform motion deblurring challenge report

    George Ciubotariu, Florin-Alexandru Vasluianu, Zhuyun Zhou, Nancy Mehta, Radu Timofte, et al. AIM 2025 high FPS non-uniform motion deblurring challenge report. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  9. [9]

    The cityscapes dataset for semantic urban scene understanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. In Proceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 3213–3223, 2016. 2

    work page 2016

  10. [10]

    A.H. Da F. Klein, G.G. Santana, F.L. Diehl, and J.T. De Menezes. Analysis of hazards associated with sea bathing: results of five years work in oceanic beaches of Santa Catarina state, southern Brazil. Journal of Coastal Re- search, pages 107–116, 2003. 2

    work page 2003

  11. [11]

    Automated rip current detection with region based convolutional neural networks

    Akila de Silva, Issei Mori, Gregory Dusek, James Davis, and Alex Pang. Automated rip current detection with region based convolutional neural networks. Coastal Engineering, 166:103859, 2021. 2, 3. 6

    work page 2021

  12. [12]

    Automated rip current detection with region based convolutional neural networks

    Akila de Silva, Issei Mori, Gregory Dusek, James Davis, and Alex Pang. Automated rip current detection with region based convolutional neural networks. Coastal Engineering, 166:103859, 2021. 2

    work page 2021

  13. [13]

    RipViz: Find- ing Rip Currents by Learning Pathline Behavior

    Akila de Silva, Mona Zhao, Donald Stewart, Fahim Hasan, Gregory Dusek, James Davis, and Alex Pang. RipViz: Find- ing Rip Currents by Learning Pathline Behavior. IEEE Transactions on Visualization and Computer Graphics , 30 (7):3930–3944, 2024. 2

    work page 2024

  14. [14]

    Rip currents in the non-tidal surf zone with sandbars: numerical analysis versus field measurements

    Aleksandra Dudkowska, Aleksandra Borun, Jakub Malicki, Jan Sch ¨onhofer, and Gabriela Gic-Grusza. Rip currents in the non-tidal surf zone with sandbars: numerical analysis versus field measurements. Oceanologia, 62(3):291–308,

    work page

  15. [15]

    Rip Current Segmentation: A novel benchmark and YOLOv8 baseline results

    Andrei Dumitriu, Florin Tatui, Florin Miron, Radu Tudor Ionescu, and Radu Timofte. Rip Current Segmentation: A novel benchmark and YOLOv8 baseline results. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 1261–1271,

    work page

  16. [16]

    AIM 2025 challenge on rip current seg- mentation (RipSeg)

    Andrei Dumitriu, Florin Miron, Florin Tatui, Radu Tudor Ionescu, Radu Timofte, Aakash Ralhan, Florin-Alexandru Vasluianu, et al. AIM 2025 challenge on rip current seg- mentation (RipSeg). In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision (ICCV) Workshops,

    work page 2025

  17. [17]

    Ripvis: Rip currents video instance segmentation benchmark for beach monitor- ing and safety

    Andrei Dumitriu, Florin Tatui, Florin Miron, Aakash Ralhan, Radu Tudor Ionescu, and Radu Timofte. Ripvis: Rip currents video instance segmentation benchmark for beach monitor- ing and safety. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3427–3437, 2025. 1, 2, 7

    work page 2025

  18. [18]

    Efficient real-world deblurring using single images: AIM 2025 chal- lenge report

    Daniel Feijoo, Paula Garrido, Marcos Conde, Jaesung Rim, Alvaro Garcia, Sunghyun Cho, Radu Timofte, et al. Efficient real-world deblurring using single images: AIM 2025 chal- lenge report. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  19. [19]

    Cubuk, Quoc V

    Golnaz Ghiasi, Yin Cui, Aravind Srinivas, Rui Qian, Tsung- Yi Lin, Ekin D. Cubuk, Quoc V . Le, and Barret Zoph. Simple copy-paste is a strong data augmentation method for instance segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2918–2928, 2021. 5

    work page 2021

  20. [20]

    Mask R-CNN

    Kaiming He, Georgia Gkioxari, Piotr Doll ´ar, and Ross Gir- shick. Mask R-CNN. In Proceedings of the IEEE Interna- tional Conference on Computer Vision (ICCV), pages 2961– 2969, 2017. 2

    work page 2017

  21. [21]

    Numerical study of rip currents inter- laced with multichannel sandbars

    Xiao Hong, Yao Zhang, Bin Wang, Shuihua Zhou, Shengbin Yu, and Juan Zhang. Numerical study of rip currents inter- laced with multichannel sandbars. Natural Hazards, 108(1): 593–605, 2021. 6

    work page 2021

  22. [22]

    Daformer: Improving network architectures and training strategies for domain-adaptive semantic segmentation

    Lars Hoyer, Dengxin Dai, and Luc Van Gool. Daformer: Improving network architectures and training strategies for domain-adaptive semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9924–9935, 2022. 3

    work page 2022

  23. [23]

    4K image super-resolution on mobile NPUs: Mobile AI & AIM 2025 challenge report

    Andrey Ignatov, Georgy Perevozchikov, Radu Timofte, et al. 4K image super-resolution on mobile NPUs: Mobile AI & AIM 2025 challenge report. InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Work- shops, 2025. 2

    work page 2025

  24. [24]

    Efficient learned smartphone ISP on mobile GPUs: Mo- bile AI & AIM 2025 challenge report

    Andrey Ignatov, Georgy Perevozchikov, Radu Timofte, et al. Efficient learned smartphone ISP on mobile GPUs: Mo- bile AI & AIM 2025 challenge report. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  25. [25]

    Efficient image denoising on smartphone GPUs: Mobile AI & AIM 2025 challenge report

    Andrey Ignatov, Georgy Perevozchikov, Radu Timofte, et al. Efficient image denoising on smartphone GPUs: Mobile AI & AIM 2025 challenge report. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  26. [26]

    Adapting stable diffusion for on-device inference: Mobile AI & AIM 2025 challenge report

    Andrey Ignatov, Georgy Perevozchikov, Radu Timofte, et al. Adapting stable diffusion for on-device inference: Mobile AI & AIM 2025 challenge report. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  27. [27]

    YOLO by Ultralytics, 2023

    Glenn Jocher, Ayush Chaurasia, and Jing Qiu. YOLO by Ultralytics, 2023. 2

    work page 2023

  28. [28]

    AIM 2025 challenge on robust offline video super-resolution: Dataset, methods and results

    Nikolai Karetin, Ivan Molodetskikh, Dmitry Vatolin, Radu Timofte, et al. AIM 2025 challenge on robust offline video super-resolution: Dataset, methods and results. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision (ICCV) Workshops, 2025. 2

    work page 2025

  29. [29]

    RipFinder: Real-time rip current detection on mobile devices

    Fahim Khan, Akila De Silva, Ashleigh Palinkas, Gregory Dusek, James Davis, and Alex Pang. RipFinder: Real-time rip current detection on mobile devices. Frontiers in Marine Science, 12:1549513, 2025. 2

    work page 2025

  30. [30]

    Rip- scout: Realtime ml-assisted rip current detection and auto- mated data collection using uavs

    Fahim Khan, Donald Stewart, Akila de Silva, Ashleigh Palinkas, Gregory Dusek, James Davis, and Alex Pang. Rip- scout: Realtime ml-assisted rip current detection and auto- mated data collection using uavs. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing ,

    work page

  31. [31]

    YOLOv11: An Overview of the Key Architectural Enhancements

    R Khanam and M Hussain. YOLOv11: An overview of the key architectural enhancements. arXiv preprint arXiv:2410.17725, 2024. 2. 4, 5

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  32. [32]

    Segment anything

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment anything. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 4015–4026,

    work page

  33. [33]

    Real-world raw de- noising using diverse cameras: AIM 2025 challenge report

    Feiran Li, Jiacheng Li, Marcos Conde, Beril Besbinar, Vlad Hosu, Daisuke Iso, Radu Timofte, et al. Real-world raw de- noising using diverse cameras: AIM 2025 challenge report. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  34. [34]

    Microsoft COCO: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft COCO: Common objects in context. In Proceedings of 13th European conference on Computer Vi- sion (ECCV), pages 740–755. Springer, 2014. 2

    work page 2014

  35. [35]

    AIM 2025 perceptual image super-resolution chal- lenge

    Bruno Longarela, Marcos Conde, ´Alvaro Garc´ıa, Radu Tim- ofte, et al. AIM 2025 perceptual image super-resolution chal- lenge. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision (ICCV) Workshops, 2025. 2

    work page 2025

  36. [36]

    James B. Lushine. A study of rip current drownings and related weather factors. National Weather Digest, 16(3):13– 19, 1991. 2

    work page 1991

  37. [37]

    Rtmdet: An empirical study of de- signing real-time object detectors,

    Chengqi Lyu, Wenwei Zhang, Haian Huang, Yue Zhou, Yudong Wang, Yanyi Liu, Shilong Zhang, and Kai Chen. RT- MDet: An Empirical Study of Designing Real-Time Object Detectors. arXiv preprint arXiv:2212.07784, 2022. 4, 5

    work page arXiv 2022

  38. [38]

    Machine learning appli- cations in detecting rip channels from images

    Corey Maryan, Md Tamjidul Hoque, Christopher Michael, Elias Ioup, and Mahdi Abdelguerfi. Machine learning appli- cations in detecting rip channels from images. Applied Soft Computing, 78:84–93, 2019. 2

    work page 2019

  39. [39]

    McGill and Jean T

    Sean P. McGill and Jean T. Ellis. Rip current and channel detection using surfcams and optical flow. Shore & Beach, 90(1):50, 2022

    work page 2022

  40. [40]

    Flow-based rip current detection and visualiza- tion

    Issei Mori, Akila de Silva, Gregory Dusek, James Davis, and Alex Pang. Flow-based rip current detection and visualiza- tion. IEEE Access, 10:6483–6495, 2022. 2

    work page 2022

  41. [41]

    What is a rip current? https://oceanservice.noaa.gov/facts/ripcurrent.html, 2023

    National Oceanic and Atmospheric Admin- istration (NOAA). What is a rip current? https://oceanservice.noaa.gov/facts/ripcurrent.html, 2023. Accessed: March, 2023. 2

    work page 2023

  42. [42]

    Detecting and Visualizing Rip Current Using Optical Flow

    Shweta Philip and Alex Pang. Detecting and Visualizing Rip Current Using Optical Flow. In Proceedings of the Euro- graphics / IEEE VGTC Conference on Visualization: Short Papers, pages 19–23, 2016. 2

    work page 2016

  43. [43]

    RipGAN: A GAN-based rip current data augmenta- tion method

    Shenyang Qian, Mitchell Harley, Imran Razzak, and Yang Song. RipGAN: A GAN-based rip current data augmenta- tion method. In Proceedings of the IEEE International Con- ference on Robotics and Automation, 2025. 3

    work page 2025

  44. [44]

    Interpretable deep learning applied to rip cur- rent detection and localization

    Neelesh Rampal, Tom Shand, Adam Wooler, and Christo Rautenbach. Interpretable deep learning applied to rip cur- rent detection and localization. Remote Sensing , 14(23): 6048, 2022. 2

    work page 2022

  45. [45]

    Ripnet: A lightweight one-class deep neural network for the identification of rip currents

    Ashraf Haroon Rashid, Imran Razzak, Muhammad Tanveer, and Antonio Robles-Kelly. Ripnet: A lightweight one-class deep neural network for the identification of rip currents. In Proceedings of 27th International Conference on Neural In- formation Processing, pages 172–179, 2020

    work page 2020

  46. [46]

    RipDet: A fast and lightweight deep neural network for rip currents detection

    Ashraf Haroon Rashid, Imran Razzak, Muhammad Tanveer, and Antonio Robles-Kelly. RipDet: A fast and lightweight deep neural network for rip currents detection. In Proceed- ings of 2021 International Joint Conference on Neural Net- works, pages 1–6, 2021

    work page 2021

  47. [47]

    Tanveer, and Michael Hobbs

    Ashraf Haroon Rashid, Imran Razzak, M. Tanveer, and Michael Hobbs. Reducing rip current drowning: An im- proved residual based lightweight deep architecture for rip detection. ISA Transactions, 132:199–207, 2023. 2

    work page 2023

  48. [48]

    SAM 2: Segment Anything in Images and Videos

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman R¨adle, Chloe Rolland, Laura Gustafson, et al. SAM 2: Segment Anything in Images and Videos. arXiv preprint arXiv:2408.00714, 2024. 2

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  49. [49]

    AIM 2025 challenge on screen-content video quality assessment: Methods and results

    Nickolay Safonov, Mikhail Rakhmanov, Dmitriy Vatolin, Radu Timofte, et al. AIM 2025 challenge on screen-content video quality assessment: Methods and results. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision (ICCV) Workshops, 2025. 2

    work page 2025

  50. [50]

    A review on YOLOv8 and its advancements

    Mupparaju Sohan, Thotakura Sai Ram, and Ch Venkata Rami Reddy. A review on YOLOv8 and its advancements. In Proceedings of the International Conference on Data Intelli- gence and Cognitive Informatics , pages 529–545. Springer,

    work page

  51. [51]

    AIM 2025 challenge on inverse tone mapping report: Methods and results

    Chao Wang, Francesco Banterle, Bin Ren, Radu Timofte, et al. AIM 2025 challenge on inverse tone mapping report: Methods and results. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision (ICCV) Workshops,

    work page 2025

  52. [52]

    CBAM: Convolutional Block Attention Module

    Sanghyun Woo, Jongchan Park, Joon-Young Lee, and In So Kweon. CBAM: Convolutional Block Attention Module. In Proceedings of the European Conference on Computer Vi- sion (ECCV), pages 3–19, 2018. 7

    work page 2018

  53. [53]

    Codabench: Flexible, easy-to-use, and reproducible meta-benchmark platform

    Zhen Xu, Sergio Escalera, Adrien Pav ˜ao, Magali Richard, Wei-Wei Tu, Quanming Yao, Huan Zhao, and Isabelle Guyon. Codabench: Flexible, easy-to-use, and reproducible meta-benchmark platform. Patterns, 3(7):100543, 2022. 2

    work page 2022

  54. [54]

    AIM 2025 low-light raw video denoising challenge: Dataset, methods and results

    Alexander Yakovenko, George Chakvetadze, Ilya Khrapov, Maksim Zhelezov, Dmitry Vatolin, Radu Timofte, et al. AIM 2025 low-light raw video denoising challenge: Dataset, methods and results. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision (ICCV) Workshops,

    work page 2025

  55. [55]

    Video instance seg- mentation

    Linjie Yang, Yuchen Fan, and Ning Xu. Video instance seg- mentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 5188–5197,

    work page

  56. [56]

    The 3rd large-scale video object segmentation challenge - video in- stance segmentation track, 2021

    Linjie Yang, Yuchen Fan, Yang Fu, and Ning Xu. The 3rd large-scale video object segmentation challenge - video in- stance segmentation track, 2021. 2

    work page 2021

  57. [57]

    Style-hallucinated dual consistency learn- ing for domain generalized semantic segmentation

    Yuyuan Zhao, Zaiyu Zhong, Nan Zhao, Nicu Sebe, and Gim Hee Lee. Style-hallucinated dual consistency learn- ing for domain generalized semantic segmentation. In Pro- ceedings of the European Conference on Computer Vision (ECCV), pages 535–552, 2022. 3

    work page 2022

  58. [58]

    YOLO-Rip: A modified lightweight net- work for Rip currents detection.Frontiers in Marine Science, 9:930478, 2022

    Daoheng Zhu, Rui Qi, Pengpeng Hu, Qianxin Su, Xue Qin, and Zhiqiang Li. YOLO-Rip: A modified lightweight net- work for Rip currents detection.Frontiers in Marine Science, 9:930478, 2022. 2

    work page 2022