Disease diagnosis system and method using multiple color models and neural network

Abstract

A system and method for learning through a neural network to input images of biological tissue in different multi-color models by using a learned neural network to diagnose a predetermined disease (e.g., prostate cancer) is disclosed. According to one aspect of the present invention, as a disease diagnosis system implemented in a system including a processor and a storage device storing a neural network and using a living body image, i.e., a slide, and the neural network, the system comprises a pre-processing module for pre-processing the slide, wherein the pre-processing module generates, foreach image block formed by dividing the slide by a predetermined size, first to K-th image information (wherein K is an integer of 2 or more) corresponding to the image block, and the i-th image information (i is any integer of 1<= i <= K) is composed of at least one channel value corresponding to the i-th color model (color model), and the first color model to the K-th color model are differentcolor models; and an image block neural network that, for each of the segmented image blocks, receives the first to N-th image information corresponding to the image block as an input layer, and outputs an image block level diagnosis result indicating whether or not a disease exists in the image block.

Description

technical field [0001] The invention relates to a disease diagnosis system and method utilizing multiple color models and neural networks. In more detail, it relates to a system and method thereof capable of diagnosing a predetermined disease (such as prostate cancer). Background technique [0002] One of the main tasks performed by a pathology or pathology department is to read images of a patient's biological body and perform a diagnosis to determine the state or symptoms of a particular disease. This diagnosis is a method that depends on the long-term skilled experience of medical personnel. [0003] In recent years, due to the development of machine learning, there are active attempts to automate tasks such as image recognition or classification by using computer systems. In particular, attempts are being made to use one type of machine learning, namely neural networks (for example, deep learning methods using Convolution neural network (CNN)) to automate diagnoses th...

AI Technical Summary

Problems solved by technology

[0008] However, in this case, depending on the characteristics of the staining reagent used for staining the biological tissue corresponding to the biological image, the color of the stained tissue may vary, which will directly affect the neural network to be learned

Images