Analysis and Research of Spinal CT Image Segmentation Based on Improved Watershed Algorithm
Yuting He, Shigang Wang, Xueshan Gao · 2023
The traditional watershed is easy to be affected and then produces over-segmentation. In order to reduce the image over-segmentation, it is necessary to reduce the irrelevant information and enhance the foreground and background contrast of the spine CT images by improving the threshold segmentation. After marking the foreground and background, the watershed algorithm works in its restricted area by modifying the original gradient function. By experimental comparative analysis, the proposed method effectively restricts the image over-segmentation and also removes the holes made by the influence of noise. At the same time, the improved algorithm has universal applicability to the spine images to process pictures for saving time and laying a good foundation for the subsequent reconstruction.