Image detection and semantic segmentation of inflammatory regions for dental images

Abstract

This invention discloses an image detection and semantic segmentation method for dental images, relating to the fields of medical image processing and computer vision. The method includes S1: acquiring a set of original dental images; performing denoising on each image and addressing brightness unevenness caused by the angle of illumination; and optimizing image boundary details through contrast enhancement technology to obtain a clear first image set. This image detection and semantic segmentation method for dental images achieves stable identification and semantic-level segmentation of inflammatory areas, effectively reducing false positives and uncertainties caused by single image analysis, and improving the objectivity, consistency, and reliability of dental inflammation diagnosis. It has high clinical application value and promotional significance.

Images