Simplified Inception Module Based Hadamard Attention Mechanism for Medical Image Classification

Medical image classification has played an important role in the medical field, and related method based on deep learning become powerful technique classification. In this article, we propose a simplified inception module Hadamard attention (SI + HA) mechanism for Specifically, new mechanism: mechanism. It improves accuracy of without greatly increasing complexity model. Meanwhile, adopt to improve utilization parameters. We use two datasets prove superiority our proposed method. BreakHis dataset, AUCs can reach 98.74%, 98.38%, 98.61% 97.67% under magnification factors 40×, 100×, 200× 400×, respectively. The accuracies 95.67%, 94.17%, 94.53% 94.12% KIMIA Path 960 99.91% 99.03%. is superior currently popular methods significantly effectiveness