Novel auxiliary diagnosis and treatment method for verruca vulgaris based on MultiResUnet

Abstract

The invention designs a novel auxiliary diagnosis and treatment method for verruca vulgaris based on MultiResUnet, which is mainly used for solving the problems that the shape of part of verruca vulgaris is small and not obvious, the number of the verruca vulgaris caused by the infectivity of the verruca vulgaris is uncertain, a doctor only judges a cryotherapy position with naked eyes, the difficulty and time consumption are high, and when the number of patients is large, the time is short. The problems that a doctor mistakenly judges the freezing position of verruca vulgaris due to exhaustion, so that other irrelevant skin tissues are hurt, and unnecessary hurt is caused to a patient are solved. In order to achieve the purpose, the adopted processing method comprises the following steps: carrying out mass training on a MultiResUnet network by using a verruca vulgaris picture data set of a specific region (relatively straight regions, such as a palm, a hand back, a sole and the like), finally obtaining a segmentation model with a convergent loss function, highest precision and a maximum MIoU value through fine tuning of hyper-parameters and the like, and inputting a common wart picture of a specific area into the network, so that an approximate area, located on the skin surface, of the bottom of the common wart can be segmented.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to the field of image segmentation. Background technique [0002] In recent years, with the continuous development of the field of deep learning, the technology of deep learning in the field of medical image processing has become more and more mature. Among them, it is widely used in the field of medical image segmentation. The FullyConvolutional Network (FCN) proposed by Long et al. performs pixel-level classification of images, thus solving the problem of image segmentation at the semantic level. Unlike the classic CNN that uses the fully connected layer to obtain fixed-length feature vectors for classification in the convolutional layer, FCN can accept input images of any size, and use the deconvolution layer to upsample the feature map of the last convolutional layer, so that It reverts to the same dimensions as the input image, allowing a prediction for every pixel, while p...

AI Technical Summary

Problems solved by technology

[0004] The present invention designs a new auxiliary diagnosis and treatment method for common warts based on MultiResUnet, which is mainly used to solve the problem that some common warts are small and not obvious and the number of them is uncertain due to their infectivity. When doctors judge the location of cryotherapy with naked eyes Difficult and time-consuming, and when the number of patients is large, the doctor may make a mistake in judging the freezing position of the common wart due to fatigue, which will damage the rest of the irrelevant skin tissue and cause unnecessary harm to the patient.

Images