Biomedical Image Processing and Applications Based on Multi Object Detection Algorithm of Computational Vision

Medical image analysis based on deep learning has important research significance for accurately locating and identifying lesion targets. This article aims to address the issues of improving the detection efficiency and performance of existing object detection methods, and introduces the SE Net module to the Yolo v4 algorithm. The new algorithm improves the positioning accuracy of the data used in this article, achieving the goal of using object detection algorithms for rough target localization before classification. We selected Faster RCNN, Yolo v4, and improved Yolo v4 algorithms for experiments on the BraTS2018 dataset. The experimental results show that, on the whole, the improved Yolo v4 algorithm has better classification effect, and the improved Yolo v4 algorithm provides a good method for assisting doctors to diagnose brain tumor diseases.