In this paper, we propose a new method for CT pathological image analysis of brain and chest to extract image features and classify images. Because the deep neural network needs a large number of ...

In the study of image classification, neural network learning relies heavily on datasets. Due to variability in the difficulty of collecting images in reality, datasets tend to have class imbalance ...