Scholarlink

Thyroid cancer is a prevalent endocrine malignancy, and its timely diagnosis is crucial for effective treatment and optimal patient outcomes. Deep learning algorithms have emerged as promising tools for medical image segmentation, facilitating accurate localization and characterization of thyroid nodules. This study investigated the application of deep learning techniques for thyroid cancer image segmentation, with the aim of improving diagnostic accuracy and refining clinical outcomes. A thyroid image dataset was classified into two categories: malignant and benign. A convolutional neural network (CNN) model was employed for thyroid cancer image segmentation using a diverse dataset of thyroid ultrasound images from a reputable medical institution. The model achieved an accuracy of 92.67 percent in delineating and segmenting thyroid nodules for validation split on 0.2. This research advances our understanding and application of deep learning in thyroid cancer analysis.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep Learning-Based Thyroid Cancer Detection and Segmentation Using Convolutional Neural Networks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Aditya Praksh,
  • Juhi Singh

摘要

Thyroid cancer is a prevalent endocrine malignancy, and its timely diagnosis is crucial for effective treatment and optimal patient outcomes. Deep learning algorithms have emerged as promising tools for medical image segmentation, facilitating accurate localization and characterization of thyroid nodules. This study investigated the application of deep learning techniques for thyroid cancer image segmentation, with the aim of improving diagnostic accuracy and refining clinical outcomes. A thyroid image dataset was classified into two categories: malignant and benign. A convolutional neural network (CNN) model was employed for thyroid cancer image segmentation using a diverse dataset of thyroid ultrasound images from a reputable medical institution. The model achieved an accuracy of 92.67 percent in delineating and segmenting thyroid nodules for validation split on 0.2. This research advances our understanding and application of deep learning in thyroid cancer analysis.