Abstract:The venous’s information show in Leukoaraiosis’s MR image is complex, on account of this characteristic, a multi-threshold venous extraction method was proposed in this paper. First, the venous detail information was enhanced by using morphological top-hat transform; Second, the image’s histogram was processed by using multilayer wavelet decomposition, then threshold processing and reconstruction were used at different levels by wavelet detail information and approximation information to extract venous gray feature; Finally, the multi-threshold segmentation method was used to obtain venous regions. The results show that this method can extract venous information automatically and quickly, furthermore, it realized the venous’s quantitative analysis. It provided a quantitative criterion for doctors’ diagnosis of venous expansion in white matter regions of Leukoaraiosis and had clinical assistant diagnosis value.