REVIEW 8 cited by
SAM Struggles in Concealed Scenes -- Empirical Study on Segment Anything
Not yet reviewed by Pith; the record is open.
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
Signed reviews
read the original abstract
Segmenting anything is a ground-breaking step toward artificial general intelligence, and the Segment Anything Model (SAM) greatly fosters the foundation models for computer vision. We could not be more excited to probe the performance traits of SAM. In particular, exploring situations in which SAM does not perform well is interesting. In this report, we choose three concealed scenes, i.e., camouflaged animals, industrial defects, and medical lesions, to evaluate SAM under unprompted settings. Our main observation is that SAM looks unskilled in concealed scenes.
Forward citations
Cited by 8 Pith papers
-
Segment Concealed Objects with Incomplete Supervision
SEE is a unified mean-teacher framework that derives SAM prompts from coarse teacher masks to generate pseudo-labels, and reports state-of-the-art results for weakly and semi-supervised concealed object segmentation.
-
Inspiring the Next Generation of Segment Anything Models: Comprehensively Evaluate SAM and SAM 2 with Diverse Prompts Towards Context-Dependent Concepts under Different Scenes
A unified evaluation of SAM and SAM 2 on 11 context-dependent concepts over 33 datasets shows box prompts dominate, SAM 2 lags SAM in some static-image settings, and both are prompt-sensitive.
-
SimCMF: A Simple Cross-modal Fine-tuning Strategy from Vision Foundation Models to Any Imaging Modality
A small convolutional adapter plus a frozen patch embedding lets SAM segment depth, thermal, polarization, HHA, and NIR images far better than training from scratch, with parameter-efficient fine-tuning matching full ...
-
Promoting SAM for Camouflaged Object Detection via Selective Key Point-based Guidance
A lightweight point-prompt generator enables frozen SAM to reach state-of-the-art camouflaged object detection on several benchmarks.
-
PGP-SAM: Prototype-Guided Prompt Learning for Efficient Few-Shot Medical Image Segmentation
Prototype-guided prompt learning lets a SAM variant reach 78.75% mean Dice on Synapse and 76.39% on a ventricle dataset using 10% of training slices, beating SAMed and other prompt-free SAM baselines.
-
COMPrompter: reconceptualized segment anything model with multiprompt network for camouflaged object detection
COMPrompter, a multiprompt SAM variant, reports top camouflaged object detection scores, but its main evaluation uses ground-truth prompts as input.
-
Self-Prompt SAM: Medical Image Segmentation via Automatic Prompt SAM Adaptation
Self-Prompt-SAM automatically generates point, box, and mask prompts for a fine-tuned SAM and reports state-of-the-art Dice scores on three medical segmentation benchmarks.
- A Holistically Point-guided Text Framework for Weakly-Supervised Camouflaged Object Detection
Discussion (0). Continue with ORCID to comment.