Automatic shadow detection in intra vascular ultrasound images using adaptive thresholding

This paper describes an automated algorithm for shadow region detection in Intra Vascular Ultrasound images, using an adaptive threshold method for threshold selection and contour approach for border detection. As shadow appears behind the calcification plaque, it makes it difficult or impossible for the dark region to process automatically around these regions. The acoustic shadow usually follows the hard plaque in IVUS images and it can distinguish calcification regions from other bright regions. Therefore we propose to use Otsu Threshold for calcification plaque segmentation and the Active contours without edge method for shadow region separation of the image. Results show that the proposed meth efficiently detected shadow regions even in complicated images. This proposed algorithm presented specificity of 86% and sensitivity of 93%.