A classification method for seborrheic keratosis and flat wart

Abstract

The invention discloses a method for classifying seborrheic keratosis and flat wart diseases. Firstly, from a small number of marked characteristic pictures and non-characteristic pictures, an incremental model that automatically increases the data set is trained, and the unmarked The pictures are automatically labeled, and then the obtained labeled feature pictures are trained to obtain a deep convolutional neural network classifier, thereby providing an efficient and fast auxiliary identification method for doctors' diagnosis.

Description

technical field [0001] The invention relates to the fields of computer and medicine, in particular to a method for classifying seborrheic keratosis and flat warts. Background technique [0002] Clinically, the diagnosis of skin diseases mainly relies on the naked eye observation and subjective experience of doctors, and lacks scientific quantification means. Seborrheic keratosis and flat warts are the most common skin diseases that affect appearance in dermatology. The distribution and clinical manifestations of these two diseases are very similar, and even experienced dermatologists are difficult to distinguish them. In addition, with the sharp increase in outpatient visits, doctors need to observe a large number of cases and read a large amount of imaging data, and their diagnostic workload is also unbearable. Therefore, it is necessary to provide a convenient and effective method for the identification of seborrheic keratosis and flat wart, which can help doctors reduce ...

AI Technical Summary

Problems solved by technology

[0005] (1) The histopathological examination in the prior art is rarely used at present, the reason is that the histopathological biopsy is traumatic and easy to leave scars, which is often difficult for patients to accept

[0006] (2) The result of dermoscopy is a partially enlarged photo, and no microstructure can be seen. In view of the extremely similar clinical manifestations of the two, the accuracy of this method is not high enough.

[0007] (3) Although confocal microscopy can obtain the microstructure of the skin, when the confocal microscope scans the lesion and the surrounding skin horizontally and vertically, it will generate a large number of images, and the proportion of characteristic and diagnostic images is not high , these massive pictures will undoubtedly increase the work of dermatologists, and there may be artificial misjudgments and missed diagnoses

Images