A segmentation processing method for effectively solving the problem of partial covering of angiography images

In this paper, an image segmentation method is proposed to deal with the partial contamination or partial covering of the region of interest. This method can be used to segment and extract the vessels or the catheter of the X-ray coronary angiography sequence image. The main steps of this method include: firstly, image preprocessing. Preprocessing includes several steps such as image filtering, Fourier transform, and image enhancement. The live images and masks of angiography images require different processing methods. For masks that are not covered by contrast agents, the routine edge detection is sufficient. However, in the case of live film, the catheter or blood vessel may be covered by the contrast agent. The segmentation method is as follows: after image preprocessing, select the pixel with the lowest gray value as the seed point to grow the region of specific elements. Using morphological processing to make the growing area more fully and connected. Based on this connected domain, it expands twice to form an expanded area. Make the extended area cover the spilled contrast agent as much as possible to determine the contrast agent range. After that, the gray value is decreased for the growth area, and the gray value is increased for the contrast medium area, and the gray difference is enlarged. Using the bottom-hat transformation, the covered part of the blood vessel or catheter is more prominent, which is convenient for segmentation. Finally, the canny operator is used for edge detection, and the largest connected domain is retained, so that a complete blood vessel or catheter can be extracted. The experiment shows that the method can effectively solve the problem of vessel or catheter segmentation and extraction in the covered film with good controllability and robustness.