Abstract: In order to acquire more accurate shape information of sublingual vein,a near infrared sublingual vein image acquisition device is employed to capture the near infrared sublingual images,and a corresponding segmentation method of sublingual vein is proposed as well.According to the characteristics of the captured images,at first a watershed-like technique is used for background removal,followed by a dynamic thresholding to binarize the detected effective sublingual region,then candidate sublingual veins regions are extracted by combining image intensity features.Finally,the intensity consistency of sublingual vein region is used to assist the adaptive region growing,and the more accurate contour of sublingual veins is obtained.Experimental results show that the proposed segmentation method can efficiently isolate the sublingual veins from the surrounding tongue proper,and accurately segment the sublingual veins.