Extraction of mask image for DSA using principal component analysis

Digital subtraction angiography (DSA) is technique for visualization of the blood vessels in the human body. And it is necessary to obtain an image without a contrast medium as a mask image. In the conventional method, this mask image is commonly selected arbitrarily from a series of angiograms, and thus DSA images are sometimes degraded by the mask image, and the mask image that is used for subtracting a background structure is very important to keep the quality of DSA images. The aim is to extract an optimum mask image from a series of acquired raw data. To make this mask image, we used the principal component analysis and we adopted the first component image as this mask image. The results obtained by this method showed the effectiveness of the proposed method. Furthermore, we have also succeeded in visualizing each blood vessel component image.