# Understanding Segment Anything Model: SAM is Biased Towards Texture Rather than Shape

arXiv: https://arxiv.org/abs/2311.11465 · HTML: https://mlsanity.com/paper/2311.11465/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 34 of its 48 references that are in the map.
- First posted on arXiv 3 Jun 2023.

## Builds on (most cited first, 30 of 34)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [Densely Connected Convolutional Networks](https://mlsanity.com/paper/1608.06993/) (2016)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [On the Opportunities and Risks of Foundation Models](https://mlsanity.com/paper/2108.07258/) (2021)
- [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://mlsanity.com/paper/2102.05918/) (2021)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness](https://mlsanity.com/paper/1811.12231/) (2018)
- [Florence: A New Foundation Model for Computer Vision](https://mlsanity.com/paper/2111.11432/) (2021)
- [Intriguing Properties of Vision Transformers](https://mlsanity.com/paper/2105.10497/) (2021)
- [Text-to-image Diffusion Models in Generative AI: A Survey](https://mlsanity.com/paper/2303.07909/) (2023)
- [Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNet](https://mlsanity.com/paper/1904.00760/) (2019)
- [Track Anything: Segment Anything Meets Videos](https://mlsanity.com/paper/2304.11968/) (2023)
- [Inpaint Anything: Segment Anything Meets Image Inpainting](https://mlsanity.com/paper/2304.06790/) (2023)
- [Can SAM Segment Anything? When SAM Meets Camouflaged Object Detection](https://mlsanity.com/paper/2304.04709/) (2023)
- [Anything-3D: Towards Single-view Anything Reconstruction in the Wild](https://mlsanity.com/paper/2304.10261/) (2023)
- [A Complete Survey on Generative AI (AIGC): Is ChatGPT from GPT-4 to GPT-5 All You Need?](https://mlsanity.com/paper/2303.11717/) (2023)
- [A Survey on Audio Diffusion Models: Text To Speech Synthesis and Enhancement in Generative AI](https://mlsanity.com/paper/2303.13336/) (2023)
- [Generative AI meets 3D: A Survey on Text-to-3D in AIGC Era](https://mlsanity.com/paper/2305.06131/) (2023)
- [A Survey on Masked Autoencoder for Self-supervised Learning in Vision and Beyond](https://mlsanity.com/paper/2208.00173/) (2022)
- [Input Augmentation with SAM: Boosting Medical Image Segmentation with Segmentation Foundation Model](https://mlsanity.com/paper/2304.11332/) (2023)
- [One Small Step for Generative AI, One Giant Leap for AGI: A Complete Survey on ChatGPT in AIGC Era](https://mlsanity.com/paper/2304.06488/) (2023)
- [Cognitive Psychology for Deep Neural Networks: A Shape Bias Case Study](https://mlsanity.com/paper/1706.08606/) (2017)
- [A Survey on Segment Anything Model (SAM): Vision Foundation Model Meets Prompt Engineering](https://mlsanity.com/paper/2306.06211/) (2023)
- [Grad-StyleSpeech: Any-speaker Adaptive Text-to-Speech Synthesis with Diffusion Models](https://mlsanity.com/paper/2211.09383/) (2022)

## Built on by (most cited first, 10 of 10)

- [DarkSAM: Fooling Segment Anything Model to Segment Nothing](https://mlsanity.com/paper/2409.17874/) (2024)
- [Robustness of SAM: Segment Anything Under Corruptions and Beyond](https://mlsanity.com/paper/2306.07713/) (2023)
- [Internal-External Boundary Attention Fusion for Glass Surface Segmentation](https://mlsanity.com/paper/2307.00212/) (2023)
- [Quantifying the Limits of Segmentation Foundation Models: Modeling Challenges in Segmenting Tree-Like and Low-Contrast Objects](https://mlsanity.com/paper/2412.04243/) (2024)
- [SAM Meets UAP: Attacking Segment Anything Model With Universal Adversarial Perturbation](https://mlsanity.com/paper/2310.12431/) (2023)
- [Red-Teaming Segment Anything Model](https://mlsanity.com/paper/2404.02067/) (2024)
- [Transferring Styles for Reduced Texture Bias and Improved Robustness in Semantic Segmentation Networks](https://mlsanity.com/paper/2507.10239/) (2025)
- [Physical Object Understanding with a Physically Controllable World Model](https://mlsanity.com/paper/2606.00439/) (2026)
- [Contour Field based Elliptical Shape Prior for the Segment Anything Model](https://mlsanity.com/paper/2504.12556/) (2025)
- [LSP-ST: Ladder Shape-Biased Side-Tuning for Robust Infrared Small Target Detection](https://mlsanity.com/paper/2504.14481/) (2025)

## Abstract (arXiv)

In contrast to the human vision that mainly depends on the shape for recognizing the objects, deep image recognition models are widely known to be biased toward texture. Recently, Meta research team has released the first foundation model for image segmentation, termed segment anything model (SAM), which has attracted significant attention. In this work, we understand SAM from the perspective of texture textitv.s. shape. Different from label-oriented recognition tasks, the SAM is trained to predict a mask for covering the object shape based on a promt. With this said, it seems self-evident that the SAM is biased towards shape. In this work, however, we reveal an interesting finding: the SAM is strongly biased towards texture-like dense features rather than shape. This intriguing finding is supported by a novel setup where we disentangle texture and shape cues and design texture-shape cue conflict for mask prediction.