Pathological image classification device and method based on depth feature fusion and use method of device

Abstract

The invention discloses a pathological image classification device based on depth feature fusion, and the device comprises an image data set obtaining unit which obtains a small image data set; the specific number of integration units are used for receiving small block images in the small block image data set, enabling the small block images to repeatedly obtain further weighted and deepened image convolution features through operation of a convolution kernel and a weighting channel and dimension deepened cascading according to set network training parameters, and outputting the image convolution features; deep image convolution features and shallow image convolution features are extracted; the feature fusion unit is used for carrying out operation on the obtained deep image convolution features and the shallow image convolution features and then carrying out cascade connection on the obtained deep image convolution features and the shallow image convolution features and vectorized deep image convolution features to obtain image depth fusion features; and the classifier is used for obtaining a classification label of the image according to the image depth fusion feature obtained by the feature fusion module and in combination with a preset classification standard. The invention further discloses a pathological image classification method based on deep feature fusion. The invention discloses a method for using a pathological image classification device based on deep feature fusion. According to the invention, more accurate pathological image classification can be realized.

Description

technical field [0001] The present invention relates to the field of pathological image classification, and more specifically, to a pathological image classification technology based on deep feature fusion. Background technique [0002] In recent years, with the development of deep learning, the pathological image classification method based on convolutional neural network has been developed to some extent, and has been applied to some computer-aided diagnosis systems. However, due to the limitation of collecting and labeling pathological image data, the amount of data is small , noise data and other issues seriously affect the accuracy and reliability of the model. [0003] Therefore, experts in the field have conducted a series of explorations and researches on pathological image classification methods based on convolutional neural networks. In 2016, H. Kallen used the convolutional features extracted by the convolutional neural network to train random forest and SVM clas...

AI Technical Summary

Problems solved by technology

The convolutional neural network used in the above method only maps the pathological images layer by layer, focusing on how to improve the accuracy of classification, but the convolution features are not fused with the pathological images, which limits the classification performance of the model and limits The classification accuracy of the model

Images