Vascular extraction based on morphological and minimum class variance

A fast threshold segmentation algorithm based on the minimum interclass variance and morphology was proposed for noise removal and target-background segmentation of the vascular images. First, the minimum interclass variance method was employed to locate partition quickly. And then morphology method was used to calculate statistics pixels for judging the noise. The theoretic analysis and experiments indicate that the presented filter algorithm suitable for vascular image extracting target, and can adaptively suppress noise. Moreover, the present filter algorithm has the higher segmentation precision and lower computation complexity, which is helpful for further target recognition.