mlsanityData to 14 Oct 2026

Segment Anything Model (SAM) Meets Glass: Mirror and Transparent Objects Cannot Be Easily Detected

arXiv 2305.00278 ↗First posted 29 Apr 2023cs.CV cs.AI cs.LGcited by 34

Not on the map. Why?

Abstract (arXiv)

Meta AI Research has recently released SAM (Segment Anything Model) which is trained on a large segmentation dataset of over 1 billion masks. As a foundation model in the field of computer vision, SAM (Segment Anything Model) has gained attention for its impressive performance in generic object segmentation. Despite its strong capability in a wide range of zero-shot transfer tasks, it remains unknown whether SAM can detect things in challenging setups like transparent objects. In this work, we perform an empirical evaluation of two glass-related challenging scenarios: mirror and transparent objects. We found that SAM often fails to detect the glass in both scenarios, which raises concern for deploying the SAM in safety-critical situations that have various forms of glass.

Key facts. Cited by 34 papers in the map (data through 14 Oct 2026). Builds on 18 of its 33 references that are in the map. First posted on arXiv 29 Apr 2023.

Builds on

18 of 33 references.

References · 18

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2023Segment Anythingcited by 8,234
  3. 2021On the Opportunities and Risks of Foundation Modelscited by 4,033
  4. 2016Pyramid Scene Parsing Networkcited by 2,479
  5. 2019PointRend: Image Segmentation as Renderingcited by 440
  6. 2023Text-to-image Diffusion Models in Generative AI: A Surveycited by 265
  7. 2017ICNet for Real-Time Semantic Segmentation on High-Resolution Imagescited by 261
  8. 2020Multi-scale Interactive Network for Salient Object Detectioncited by 212
  9. 2016Deeply supervised salient object detection with short connectionscited by 196
  10. 2020Context Prior for Scene Segmentationcited by 121
  11. 2018Graph-Based Global Reasoning Networkscited by 120
  12. 2018Reverse Attention for Salient Object Detectioncited by 113
6 more
  1. 2020Segmenting Transparent Objects in the Wildcited by 98
  2. 2023A Survey on Audio Diffusion Models: Text To Speech Synthesis and Enhancement in Generative AIcited by 68
  3. 2017Direction-aware Spatial Context Features for Shadow Detectioncited by 56
  4. 2023One Small Step for Generative AI, One Giant Leap for AGI: A Complete Survey on ChatGPT in AIGC Eracited by 55
  5. 2021Enhanced Boundary Learning for Glass-like Object Segmentationcited by 48
  6. 2023A Survey on Graph Diffusion Models: Generative AI in Science for Molecule, Protein and Materialcited by 35

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 19; 2024: 9; 2025: 4; 2026: 2.

34 papers.

Sort:

By area:

  1. 2023Faster Segment Anything: Towards Lightweight SAM for Mobile Applicationscited by 352
  2. 2023EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anythingcited by 188
  3. 2023SAM-6D: Segment Anything Model Meets Zero-Shot 6D Object Pose Estimationcited by 117
  4. 2023Weakly-Supervised Concealed Object Segmentation with SAM-based Pseudo Labeling and Multi-scale Feature Groupingcited by 92
  5. 2023A Comprehensive Survey on Segment Anything Model for Vision and Beyondcited by 64
  6. 2023A Survey on Segment Anything Model (SAM): Vision Foundation Model Meets Prompt Engineeringcited by 45
  7. 2023MobileSAMv2: Faster Segment Anything to Everythingcited by 42
  8. 2024RobustSAM: Segment Anything Robustly on Degraded Imagescited by 36
  9. 2023Robustness of SAM: Segment Anything Under Corruptions and Beyondcited by 31
  10. 2023Improving the Generalization of Segmentation Foundation Model under Distribution Shift via Weakly Supervised Adaptationcited by 31
  11. 2024Efficient Track Anythingcited by 21
  12. 2025SAMRefiner: Taming Segment Anything Model for Universal Mask Refinementcited by 20
22 more
  1. 2023Black-box Targeted Adversarial Attack on Segment Anything (SAM)cited by 13
  2. 2023Open-NeRF: Towards Open Vocabulary NeRF Decompositioncited by 13
  3. 2024ASAM: Boosting Segment Anything Model with Adversarial Tuningcited by 12
  4. 2023Understanding Segment Anything Model: SAM is Biased Towards Texture Rather than Shapecited by 10
  5. 2024PosSAM: Panoptic Open-vocabulary Segment Anythingcited by 9
  6. 2023Segment Anything Meets Semantic Communicationcited by 8
  7. 2025AoP-SAM: Automation of Prompts for Efficient Segmentationcited by 7
  8. 2023Internal-External Boundary Attention Fusion for Glass Surface Segmentationcited by 6
  9. 2023Attack-SAM: Towards Attacking Segment Anything Model With Adversarial Examplescited by 5
  10. 2023AdapterShadow: Adapting Segment Anything Model for Shadow Detectioncited by 5
  11. 2023SAM Meets UAP: Attacking Segment Anything Model With Universal Adversarial Perturbationcited by 4
  12. 2023GEM: Boost Simple Network for Glass Surface Segmentation via Vision Foundation Modelscited by 3
  13. 2024Glass Segmentation with Multi Scales and Primary Prediction Guidingcited by 2
  14. 2024Segment Any Object Model (SAOM): Real-to-Simulation Fine-Tuning Strategy for Multi-Class Multi-Instance Segmentationcited by 2
  15. 2024ClassWise-SAM-Adapter: Parameter Efficient Fine-tuning Adapts Segment Anything to SAR Domain for Semantic Segmentationcited by 1
  16. 2024Urban Waterlogging Detection: A Challenging Benchmark and Large-Small Model Co-Adaptercited by 1
  17. 2023When ChatGPT for Computer Vision Will Come? From 2D to 3Dcited by 0
  18. 2024GEM: Boost Simple Network for Glass Surface Segmentation via Segment Anything Model and Data Synthesiscited by 0
  19. 2025BioDet: Boosting Industrial Object Detection with Image Preprocessing Strategiescited by 0
  20. 2025Closed-Loop Robotic Manipulation of Transparent Substrates for Self-Driving Laboratories using Deep Learning Micro-Error Correctioncited by 0
  21. 2026Glass Segmentation with Fusion of Learned and General Visual Featurescited by 0
  22. 2026Global-Local Feature Decoding with Adapter-Guided SAMv2 for Salient Object Detectioncited by 0