Otoscope image auxiliary diagnosis method based on fine-grained classification

The invention discloses an otoscope image auxiliary diagnosis method based on fine-grained classification, and the method comprises the following steps: inputting an image, obtaining a global classification result through employing a global classifier, and displaying a saliency map of a main region; selecting k local images where the lesions are located through a local selector by using the saliency map, inputting the k local images into a local classifier to obtain a classification result corresponding to each local image, and averaging the classification results to obtain a local classification result; averaging the two classification results to obtain a final diagnosis result of the image. According to the method, only the images in the training set are required to be labeled with category labels, labeling with positioning information is not required, and compared with a fine-grained classification method with strong supervision, the data collection and labeling process is simplified. According to the meth