2026· Proceedings of the 4th International Conference on Image, Algorithms, and Artificial Intelligence, ICIAAI 2026, 22-24 May 2026, Singapore, Singapore· 0 citations· 5 references
TL;DR
The review of visual foundation models represented by CLIP and SAM shows that these models significantly improve data efficiency and cross-domain generalization capabilities, but still need further exploration in medical specificity adaptation and fine-grained segmentation.
Abstract
. Medical image segmentation is a key technology for achieving precise medical care. Traditional methods rely on a large amount of labeled data and have limited generalization capabilities. Visual foundation models represented by CLIP and SAM have obtained general capabilities through large-scale pre-training, providing a new paradigm for medical segmentation. CLIP achieves semantic-guided segmentation through image-text alignment, while SAM achieves general segmentation through prompt interaction. Both can effectively reduce the reliance on labeled data. This paper systematically reviews the applications of these two types of models in medical segmentation: first, analyzes the necessity of their applications; second, separately summarizes the technical routes and adaptation progress of semantic-guided methods based on CLIP and prompt-interaction methods based on SAM; finally, discusses their advantages and challenges. The review shows that these models significantly improve data efficiency and cross-domain generalization capabilities, but still need further exploration in medical specificity adaptation and fine-grained segmentation.
A Concept-Guided Segmentation Model is proposed that integrates LLM-generated and clinically reviewed concepts into the segmentation process, and a Concept-Visual Alignment Module (CVAM) is designed to activate relevant tokens within the concepts that align with visual features, enhancing the interaction between textua...
Chang-Heng Lin, Wen-Jie Zhang, Yu-Shan Lu et al.· 0 citations
A novel IIS framework based on Multi-Layer Perceptron (MLP), which fuses clicks from users with feature representations extracted from the Transformer Encoder, which closes the gap between interactive user instructions and deep learning and provides a practical and flexible solution for accurate segmentation in many ap...
S. Raghavendra, Rithika Shyam Kumar, S. Abhilash et al.· IEEE Access· 0 citations
Medical image segmentation plays a critical role in computer-assisted diagnosis, treatment planning, surgical navigation, and quantitative medical image analysis. Among deep learning-based segmentation architectures, U-Net has become one of the most influential frameworks because of its encoder–decoder structure and sk...
Tian-Hui Wu, Shamsher Ullah, Bo-Chen Yu et al.· Journal of Imaging· 0 citations
A dual-domain structural guidance framework that integrates grayscale-guided edge features and graph-based multi-scale fusion to improve segmentation robustness and provides a reliable visual computing solution for clinical endoscopic image analysis.
Cheng-Xiang Zhang, Liang Jiang, Zhi-Chao Xu et al.· Journal of imaging informati...· 0 citations
This work proposes a conditional Flow Matching (CFM) based framework for efficient 3D medical image segmentation that maintains competitive segmentation accuracy while achieving an inference time of only 1.14 seconds per volume, offering a clear efficiency advantage over existing generative segmentation methods.
Y. Yilihamu, Jian Xue, Chuang Jia et al.· International Conference on...· 0 citations
A Prototypes Filtering and Transformation Network (PFT-Net) to alleviate the intra-class variation problem and integrates the optimal transport algorithm to convert initially filtered support prototypes into query-adapted representations, allowing them to better align with the feature distribution of the query image.
Rui-Fa Zhou, Guo-Heng Huang, Lin-Hui Wu et al.· International Conference on...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.