Scholarlink

This study aims to improve the classification of adenomyosis, a medical condition characterized by the invasion of endometrium into the myometrium, using reinforcement learning (RL) and image segmentation. By applying RL, a form of machine learning, the uterine regions of women can be classified as either malignant or benign, thereby enhancing the detection accuracy of adenomyosis. Preprocessing procedures are conducted prior to the classification phase to ensure precise quantification of adenomyosis severity on MR scans. The proposed method utilizes fuzzy clustering and adaptive neighborhood range in RL to identify regions of interest in the uterus. Evaluation metrics such as signal-to-noise ratio (SNR), peak signal-to-noise ratio (PSNR), mean square error (MSE), structural similarity index (SSIM), and Dice coefficient (DC) are used to assess the effectiveness of the approach. The results demonstrate the potential of RL-based segmentation for improving the classification and detection of adenomyosis by providing 95% accuracy, offering implications for accurate diagnosis and treatment planning for affected women.

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

Adenomyosis Segmentation Leveraging Reinforcement Learning Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Kayalvizhi Subramanian,
  • Gunasekar Thangarasu,
  • Nattar Kannan Kalliappan

摘要

This study aims to improve the classification of adenomyosis, a medical condition characterized by the invasion of endometrium into the myometrium, using reinforcement learning (RL) and image segmentation. By applying RL, a form of machine learning, the uterine regions of women can be classified as either malignant or benign, thereby enhancing the detection accuracy of adenomyosis. Preprocessing procedures are conducted prior to the classification phase to ensure precise quantification of adenomyosis severity on MR scans. The proposed method utilizes fuzzy clustering and adaptive neighborhood range in RL to identify regions of interest in the uterus. Evaluation metrics such as signal-to-noise ratio (SNR), peak signal-to-noise ratio (PSNR), mean square error (MSE), structural similarity index (SSIM), and Dice coefficient (DC) are used to assess the effectiveness of the approach. The results demonstrate the potential of RL-based segmentation for improving the classification and detection of adenomyosis by providing 95% accuracy, offering implications for accurate diagnosis and treatment planning for affected women.