Dissertation
Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

Technology of Blood Vessel Diameter Measurement Automatic Based on Digital Image Processing

Author MaoXiaoHui
Tutor ZhengLiYing;LiXiaoLei
School Harbin Engineering University
Course Computer technology
Keywords Blood vessel image Image enhancement Mathematical morphology Blood vessel diameter Pixel interval
CLC R310
Type Master's thesis
Year 2011
Downloads 15
Quotes 0
Download Dissertation

Blood vessel is an important basis for diagnosing many diseases especially vascular disease.It can reflect lesion in certain extent.Vascual diameter measurements has a widespread application in medical image processing. Because artificial visual has some shortcomings such as low efficiency,poor repeatability and comparability, measuring blood vessel has become the development trend and research focus today.This paper achieved the measurement of blood vessel automatically using digital image processing technology.In order to reduce the noise interference caused by lesion and other factors in the vascular image,this paper preprocessed the color image first,including gray-scale conversion,image filtering enhancement,thresholding and mathematical morphology.Then the direction of vessel is determined using the center line based on skeletons extraction and the key points is marked to find the vascular segments.Finally, the automatic measurement of blood vessel is achieved based on binary vessel image.The proposed algorithm has some features as follows:1) According to the gray distribution of the vessel,the paper applied Gaussian filtering and histogram normalized to enhance the blood vessel and improve the contrast of the image.In this way,the precision of the vessel segmentation can be greatly improved;2) Extracting skeleton line of blood vessel is implemented through mathematical morphology first,which cannot produce desired effect,so we extend Zhang’s fast parallel thinning algorithm using template matching to ensure the singular pixel width and connectivity;3)The algorithm of key points selection and the crossing point and endpoint recognition is proposed;4)The pixel interval is obtained by projection algorithm in space domain,which makes the blood vessel measurement achieve a higher precision.The experimental result shows that the proposed method has achieved better effect in the speed,accuracy and robustness.

Related Dissertations
More Dissertations