Pith. sign in

REVIEW 3 cited by

Evaluation of Segment Anything Model 2: The Role of SAM2 in the Underwater Environment

Not yet reviewed by Pith; the record is open.

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

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2408.02924 v1 pith:KWIBFUOC submitted 2024-08-06 cs.CV

Shijie Lian , Hua Li This is my paper · ORCID
classification cs.CV
keywords sam2underwatermodelsegmentanythingsegmentationdomainevaluation
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

With breakthroughs in large-scale modeling, the Segment Anything Model (SAM) and its extensions have been attempted for applications in various underwater visualization tasks in marine sciences, and have had a significant impact on the academic community. Recently, Meta has further developed the Segment Anything Model 2 (SAM2), which significantly improves running speed and segmentation accuracy compared to its predecessor. This report aims to explore the potential of SAM2 in marine science by evaluating it on the underwater instance segmentation benchmark datasets UIIS and USIS10K. The experiments show that the performance of SAM2 is extremely dependent on the type of user-provided prompts. When using the ground truth bounding box as prompt, SAM2 performed excellently in the underwater instance segmentation domain. However, when running in automatic mode, SAM2's ability with point prompts to sense and segment underwater instances is significantly degraded. It is hoped that this paper will inspire researchers to further explore the SAM model family in the underwater domain. The results and evaluation codes in this paper are available at https://github.com/LiamLian0727/UnderwaterSAM2Eval.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Deep Learning for Accurate Vision-based Catch Composition in Tropical Tuna Purse Seiners

    cs.CV 2025-11 conditional novelty 5.0 of 10

    YOLOv9+SAM2 segmentation with hierarchical classification estimates tuna catch composition from EM video with about 4.5% mean absolute error on controlled test operations.

  2. AquaChat: An LLM-Guided ROV Framework for Adaptive Inspection of Aquaculture Net Pens

    cs.RO 2025-07 conditional novelty 3.0 of 10

    AquaChat translates natural-language commands into symbolic ROV plans executed by a PID controller, with experiments in Gazebo and a pool; the framework runs, but several headline claims are not directly measured.

  3. A Review of Generative AI in Aquaculture: Foundations, Applications, and Future Directions for Smart and Sustainable Farming

    cs.RO 2025-07 conditional novelty 3.0 of 10

    A review that maps generative AI to aquaculture tasks, with a marine robotics case study, but the synthesis is weakened by overstated claims and weak citation support.

Pith tools