Research on Automatic Recognition of Nodule Boundaries in Thyroid Ultrasound Images based on Deep Learning

Abstract

This article provides an in-depth analysis of the application of deep learning in automatic recognition of nodule boundaries in thyroid ultrasound images, and conducts a comprehensive comparative analysis of various deep learning architectures. This study uses Mobile
U-Net as the core model to achieve specific goals, and conducts in-depth analysis of the models construction framework, parameter configuration, and training process. We have explored the specific properties and manifestations of thyroid ultrasound images, developed a comprehensive data preprocessing and post-processing operation system, and used rigorous evaluation systems and scientific experimental methods.
Empirical research has confirmed that the method proposed in this study exhibits excellent accuracy and stability performance for thyroid
nodule boundary recognition. This technical means provides reliable technical assistance for the auxiliary diagnosis of thyroid nodules, and
the promotion of intelligent diagnosis of thyroid diseases has extremely important practical significance.