Computer Vision Based Vessel Seam Detection And Tracking In Fetoscopic Images

In Twin to twin transfusion syndrome in monochrome twin pregnancies, fetus communicated through Arterio arteries, Veno Venus, Artery to Vein seam. During fetoscopic laser occlusion surgery, the identification of artery and vein separation hasthe misperception due to its color resemblance. In digital fetoscopic images the artery occurs in the bright red region, vein occurs in dark red region. Artery to vein communicative blood vessel and it is correlation regions are needed for surgery. Manually identification of these vessels is highly complicated when the laser beam is passed through the vessels. To overcome this problem. Color region based Vector Quantization method is proposed to identify the artery to vein junction. This method differentiate regions based on the colour resemblance. In this proposed method, Artery to artery (AA),Vein to Vein (VV),Artery to vein(AV) region based samples are taken to distinguish AV anastomos and coagulation. Various state of fetoscopic images are analysed based on different radiation conditions. Proposed method separates AA, VV & AV regions. Automated vessel detection and separation of different vessel regions were achieved using a system based on fetoscopic laser applied images. The automated system provides a clinically feasible and supportive method during fetus surgery. This method may be improved further for computer- or robot-assisted applications.