Efficient Semantic Segmentation Method of Lesion Area in Pathology Images Based on Uncertainty Mechanism
With the digitization of pathological images and advancements in deep learning technology,using deep learning to process high-resolution digital pathology images is crucial for expediting diagnosis and enhancing clinical efficiency.Traditional deep learning networks and image processing methods are often resource-intensive,given the high resolution of these images.This paper introduces a novel approach,combining a sliding window mechanism with multi-scale information and an uncertainty mechanism,significantly improving efficiency.This innovation reduces average processing time to within 1 min.
万方数据
维普
