Medical Image Segmentation with Bilateral Spatial Attention and Transfer Learning

Abstract

In the medical field, image processing technology is becoming more and more extensive, especially in disease screening and diagnosis. However, traditional medical image processing methods can produce a lot of noise and artifacts in low-dose X-ray CT images, affecting a doctor’s diagnostic accuracy. At the same time,  early lesions, such as colon polyps, is difficult, and existing image segmentation algorithms struggle to locate and separate lesion areas accurately. With the development of deep learning technology, especially the application of attention mechanisms in medical image processing, new solutions are provided. Attention mechanisms have shown excellent performance in noise suppression and the accurate segmentation of lesion tissue, but they rely on a large amount of training data. The lack of medical image data limits their application. In order to solve these problems, this paper proposes a medical image denoising and segmentation algorithm that combines transfer learning and attention mechanisms. We design an integrated medical image auxiliary diagnosis system based on this algorithm to improve the efficiency and accuracy of medical image processing.