李雁,邹建中,王华,冯玉洁,钟珊珊,欧阳挺.基于二维超声图像纹理分析判断HIFU凝固性坏死[J].中国医学影像技术,2010,26(6):1147~1150
基于二维超声图像纹理分析判断HIFU凝固性坏死
Analysis of coagulative necrosis caused by high-intensity focused ultrasoundwith two-dimensional ultrasonic image texture
投稿时间:2009-11-28  修订日期:2010-03-16
DOI:
中文关键词:  高强度聚焦超声消融  坏死  小波变换  支撑矢量机
英文关键词:High-intensity focused ultrasound ablation  Necrosis  Wavelet transform  Support vector machine
基金项目:国家自然科学基金面上项目(30970767)。
作者单位E-mail
李雁 重庆医科大学生物医学工程系 超声医学工程重庆市市级重点实验室,重庆 400016  
邹建中 重庆医科大学生物医学工程系 超声医学工程重庆市市级重点实验室,重庆 400016 zoujz@haifu.com.cn 
王华 重庆医科大学生物医学工程系 超声医学工程重庆市市级重点实验室,重庆 400016  
冯玉洁 重庆医科大学生物医学工程系 超声医学工程重庆市市级重点实验室,重庆 400016  
钟珊珊 重庆医科大学生物医学工程系 超声医学工程重庆市市级重点实验室,重庆 400016  
欧阳挺 重庆医科大学生物医学工程系 超声医学工程重庆市市级重点实验室,重庆 400016  
摘要点击次数: 2230
全文下载次数: 726
中文摘要:
       目的 探讨在高强度聚焦超声(HIFU)治疗中,利用二维超声监控图像在多分辨率下的纹理参数,对HIFU所致凝固性坏死组织进行评价,提高对凝固性坏死判断的敏感度。方法 在相同声强、功率和深度条件下,选用点打的方式辐照新鲜离体牛肝,采集辐照前以及辐照后即刻、1 min、2 min和3 min的二维声像图和灰度图像,利用小波变换提取二维超声图像在多分辨率下的纹理参数,建立支撑矢量机(SVM),对样本进行分析判断。结果 多分辨率下的纹理参数比灰度对HIFU凝固性坏死的判断敏感度要高,且差异有统计学意义(P<0.05)。结论 利用多分辨率下的纹理参数来评价HIFU凝固性坏死的方法是可行的,且敏感度高于灰度对凝固性坏死的评价。
英文摘要:
      Objective To assess the coagulative necrosis caused by high-intensity focused ultrasound (HIFU) with two-dimensional ultrasonic image texture, in order to improve the diagnostic sensitivity. Methods The fresh ex vivo ox livers were dot-exposed with HIFU with constant intensity, power and depth. The ultrasonography prior to exposure, the instantaneous, 1 min, 2 min and 3 min after exposure were inspected and the gray values were measured. Multi-resolution was analyzed for ultrasonic images to obtain the texture parameters under various resolutions. Support vector machine (SVM) was applied to analyze those parameters. Results The level of sensitivity for analyzing the necrosis was improved by using the texture of images, specifically the multi-resolution was analyzed to obtain the texture features from the images. Significant difference was found between different resolutions (P<0.05). Conclusion It is possible to analyze the coagulative necrosis through the texture of images. With multi-resolution analysis, the texture features from the images can be obtained and the sensitivity is higher than that of the gray values.
查看全文  查看/发表评论  下载PDF阅读器