摘要
提出了一种基于背景估计建模和精细血管提取的微脉瘤检测方法。候选微脉瘤对象由背景估计和Mahalanobis距离进行定位,非血管抑制算子结合多尺度和多滞后阈值技术实现精细的血管结构提取,经过形状分析和双环滤波处理后,非微脉瘤像素点被去除。实验结果表明,本文方法性能优于或逼近于其它同类方法,而且能够极大提高邻近血管的微脉瘤检测精度。
A microaneurysm detection approach based on background estimation modeling and finer vessel extraction is proposed. Candidate object of microaneurysm can be located by background estimation and Mahalanobis distance. A non vessel inhibition operator combining with multi-scale and multi-hysteresis threshold technique are introduced to achieve finer vasculature. And then on the basis of shape analysis and double ring filtering, non microaneurysm pixels are removed. The experiment results show that the performance of the proposed approach is better than or approximate to that of other similar approaches, and the detection accuracy of the microaneurysm closed to blood vessel is greatly improved.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2013年第2期455-461,共7页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(60835004)
湖南省科技厅计划项目(2012FJ3113)
湖南省教育厅青年项目(10B109)
湖南省重点学科建设项目资助
关键词
背景估计
微脉瘤
血管提取
糖尿病性视网膜病变
background estimation
microaneurysm
vessel extraction
diabetic retinopathy
