基于注意力机制和深度学习的颅脑外伤患者CT图像分割
CTImage Segmentation of Traumatic Brain Injury Patients Based on Attention Mechanism and Deep Learning
尹红云 1张丽娜 1王佳明 1周秀珍1
作者信息
- 1. 新疆医科大学第一附属医院神经外科,新疆乌鲁木齐 830000
- 折叠
摘要
CT脑组织图像分割对颅脑外伤的临床诊断与治疗具有重要辅助作用.基于此,研究引入基于深度学习的V-Net模型进行脑组织定位,同时引入注意力机制,以实现脑组织图像的精准分割.结果表明,研究所提分割模型的Dice指标最高达到99.81％.同时,该分割模型的精确率与召回率最高分别达到99.38％、99.84％.说明,研究所提算法具有显著的性能优势,且具有良好的实际应用效果,可为颅脑外伤的脑部诊断及治疗提供可靠的技术支持.
Abstract
CT brain tissue image segmentation plays an important auxiliary role in the clinical diagnosis and treat-ment of patients with craniocerebral trauma.Based on this,the study introduces a deep-learning-based V-Net model for brain tissue positioning,and also introduces an attention mechanism to realize accurate segmentation of brain tissue images.The results show that the Dice index of the segmentation model reached 99.81％.Meanwhile,the highest precision rate and recall rate of the segmentation model reached 99.38％and 99.84％,respectively.It shows that the proposed algorithm has significant performance advantages and good practical application effect,and provides reliable technical support for brain diagnosis and treatment of patients with brain trauma.
关键词
V-Net/CT图像分割/注意力机制/颅脑外伤/脑组织Key words
V-Net/CT image segmentation/Attention mechanism/Craniocerebral injury/Brain tissue引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据