Training deep-learning segmentation models from severely limited data

Jiahan Zhang, Yilin Song, and Steve B. Jiang.

Introduction: Deep learning has shown great potential in medical image segmentation. However, deep learning models require large amounts of annotated data to achieve good performance. In many cases, such data is not available due to privacy concerns or the high cost of annotation.

Methods: The authors used 30 head and neck computed tomography (CT) scans with well-defined contours that were deformably registered to 200 CT scans of the same anatomic site without contours. They also used a simple yet effective approach to create synthetic CT scans with contours for training deep learning segmentation models. They demonstrated its efficiency by developing auto-segmentation models for the parotids and submandibular glands. The authors show that their approach can train high-quality deep learning segmentation models from severely limited contoured cases (e.g., ~10 cases). The authors used 300, 600, 1000, and 2000 synthetic CT scans and contours generated from one PCA model to train V-Net, a 3D convolutional neural network architecture, to segment parotid and submandibular glands.

Results: The authors demonstrated that their approach can train high-quality deep learning segmentation models from severely limited contoured cases (e.g., ~10 cases). The authors used 300, 600, 1000, and 2000 synthetic CT scans and contours generated from one PCA model to train V-Net, a 3D convolutional neural network architecture, to segment parotid and submandibular glands.

Link to Article: Training deep‐learning segmentation models from severely limited data - Zhao - 2021 - Medical Physics - Wiley Online Library

This article was summarized by an AI tool that uses natural language processing. The tool is not perfect and may make mistakes or produce inaccurate or irrelevant information, but is reviewed by the post’s author prior to publishing. If you want to learn more about the article, please refer to the original source that is cited at the end of the article.