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.
维普
万方数据
