视网膜血管ssegmentation基于水平集和区域生长

资源描述视网膜lvessels发挥进口特罗勒在诊断过程fretinopathy啊。准确的分割视网膜的病理分析是至关重要的。本文weproposeanewretinalvessel首先，aretinalvesselimageis methodbasedonlevelsetandregiongrowing分割。preprocessedbythecontrast-limitedadaptivehistogramequalizationanda2dgaborwaveletto提高船舶。然后，各向异性扩散滤波器是用来smooththeimageandpreservevessel边界。最后，theregiongrowingmethodandaregion basedactivecontourmodelwithlevelset实施areappliedtoextractretinalvessels，andtheirresultsarecombinedtoachievethe最终分割。comparisonsareconductedonthepubliclyavailabledriveandstaredatabasesusing三differentmeasurements。experimentalresultsshowthattheproposedmethodreachesanaverageof94.77 % % onthestaredatabase onthedrivedatabaseand95.09精度。

17January2014-elsevier-47(2014)2437–2446.pdf