[1]高鸣△,卫元元,张博,等.基于BP-遗传算法优化的超声肿块区域分割技术*[J].生物医学工程研究,2019,02:181-185.
GAO Ming,WEI Yuanyuan,ZHANG Bo,et al.Region segmentation of ultrasonic masses based on BP-genetic algorithm[J].Journal of Biomedical Engineering Research,2019,02:181-185.
点击复制
基于BP-遗传算法优化的超声肿块区域分割技术*(PDF)
《生物医学工程研究》[ISSN:1006-6977/CN:61-1281/TN]
- 期数:
-
2019年02期
- 页码:
-
181-185
- 栏目:
-
- 出版日期:
-
2019-06-25
文章信息/Info
- Title:
-
Region segmentation of ultrasonic masses based on BP-genetic algorithm
- 文章编号:
-
1672-6278 (2019)02-0181-05
- 作者:
-
高鸣1△; 卫元元2; 张博3; 王芳4
-
1.漯河市中心医院(漯河市医专一附院)超声科,河南 漯河 462000;2. 漯河市中心医院(漯河市医专一附院)CT室,河南 漯河 462000;3. 漯河市中心医院(漯河市医专一附院)一分院内科,河南 漯河 462000;4.河南理工大学,河南焦作454150
- Author(s):
-
GAO Ming1; WEI Yuanyuan2; ZHANG Bo3; WANG Fang4
-
1.Department of Ultrasound, Luohe Central Hospital (First Affiliated Hospital of Luohe Medical College), Luohe 462000,China;2. Computerized Tomography Room,Luohe Central Hospital(First Affiliated Hospital of Luohe Medical College),Luohe 462000;3. Department of Internal Medicine, Luohe Central Hospital (First Affiliated Hospital of Luohe Medical College), Luohe 462000;4. Henan University of Technology,Jiaozuo 454150, China
-
- 关键词:
-
BP-遗传算法; 超声肿块; 区域分割; 图像; 区域模板匹配; 自适应模板; 融合
- Keywords:
-
BP- genetic algorithm; Ultrasonic masses; Region segmentation; Image; Regional template matching; Adaptive template; Integration
- 分类号:
-
R318;TP399
- DOI:
-
10.19529/j.cnki.1672-6278.2019.02.11
- 文献标识码:
-
A
- 摘要:
-
通过超声肿块区域分割处理,提高肿块检测诊断能力。提出一种基于BP-遗传算法优化的超声肿块区域分割技术。采用超声成像技术进行肿块图像采集,对采集的超声肿块图像进行块区域模板匹配处理,构建超声肿块区域检测模型,采用自适应模板特征匹配方法进行超声肿块图像融合处理,提取超声肿块区域图像的超像素特征量,根据像素特征差异度匹配方法实现超声肿块图像的关联相似度分解,以显著性特征点为中心进行超声肿块图像的区域重构,采用BP-遗传算法进行图像区域分割的自适应学习,实现超声肿块图像的高分辨辨识和分割。仿真结果表明,采用该方法进行超声肿块区域分割的精度较高,图像特征匹配性能较好,肿块区域的辨识度较高。
- Abstract:
-
In order to improve the ability of mass detection and diagnosis, an ultrasonic mass region segmentation technique based on BP-genetic algorithm was proposed. The ultrasonic imaging technique was used to collect the mass image, the block region template matching process was applied to the collected ultrasonic mass image, and the ultrasonic mass region detection model was constructed. The adaptive template feature matching method was used to fuse the ultrasonic mass image, the super-pixel feature quantity of the ultrasonic mass image was extracted, and the similarity decomposition of the ultrasonic mass image was realized according to the pixel feature difference matching method. The region reconstruction of ultrasonic mass image was carried out with the salient feature point as the center, the adaptive learning of image region segmentation using BP- genetic algorithm was adopted to realize the high resolution identification and segmentation of ultrasonic mass image. The simulation results show that the proposed method has higher accuracy, better feature matching performance and better identification degree.
备注/Memo
- 备注/Memo:
-
河南省卫计委基金资助项目(2018KY118)。△通信作者Email:m13939519288@163.com
更新日期/Last Update:
2019-07-18