MIU-Net: MIX-Attention and Inception U-Net for Histopathology Image Nuclei Segmentation

In the medical field, hematoxylin and eosin (H&E)-stained histopathology images of cell nuclei analysis represent an important measure for cancer diagnosis. The most valuable aspect is segmentation different morphologies organs subsequent diagnosis type severity disease based on pathology. recent years, deep learning techniques have been widely used in digital analysis. Automated nuclear technology enables rapid efficient tens thousands complex variable images. However, a challenging problem during blocking nuclei, overlapping, background complexity tissue fraction. To address this challenge, we present MIU-net, network structure Our proposed includes two blocks with modified inception module attention module. advantage to balance computation performance deeper layers network, combined convolutional layer using sizes kernels learn effective features fast manner complete kernel segmentation. allows us extract small fine irregular boundary from images, which can better segment cells that appear disorganized fragmented. We test our methodology public kumar datasets achieve highest AUC score 0.92. experimental results show method achieves than other state-of-the-art methods.