Advances in Semi-Supervised Medical Image Segmentation

The field of medical image segmentation is moving towards semi-supervised learning approaches, which aim to alleviate the challenge of relying on large amounts of labeled data. Researchers are exploring innovative methods to improve the accuracy and robustness of these models, including the use of uncertainty-guided pseudo-labeling, self-supervised contrastive learning, and transfer learning from 2D natural images. These advancements have the potential to significantly improve the performance of medical image segmentation models in real-world situations. Noteworthy papers include: Enhancing Dual Network Based Semi-Supervised Medical Image Segmentation with Uncertainty-Guided Pseudo-Labeling, which proposes a novel semi-supervised 3D medical image segmentation framework. Semi-Supervised 3D Medical Segmentation from 2D Natural Images Pretrained Model, which explores the transfer of knowledge from general vision models pretrained on 2D natural images to improve 3D medical image segmentation.

Sources

Enhancing Dual Network Based Semi-Supervised Medical Image Segmentation with Uncertainty-Guided Pseudo-Labeling

Transplant-Ready? Evaluating AI Lung Segmentation Models in Candidates with Severe Lung Disease

Limitations of Public Chest Radiography Datasets for Artificial Intelligence: Label Quality, Domain Shift, Bias and Evaluation Challenges

Semi-Supervised 3D Medical Segmentation from 2D Natural Images Pretrained Model

Built with on top of