Cancer cell recognition based on improved U-net convolution neural network model

Abstract

The invention discloses a method based on improved U-net convolution neural network model for cancer cell detection. The method comprises the following steps: acquiring a medical image of cancer cells, processing the medical image, labeling the cancer cells in the medical image to form a labeled data set, and dividing the data set into a training set, a verification set and a test set; Building U-Net Convolution Neural Network Model, Determine parameters of convolutional neural network model, test set and verification set are loaded into the U-net convolution neural network model, and the features are learned from the images by depth learning method, and the trained U-net convolution neural network model; The trained U-net convolution neural network model is deployed to automatically detect cancer cells in the test set The invention utilizes the bottom feature to directly learn the feature from the image, and has high running efficiency and automatic detection accuracy.

Description

technical field [0001] The invention relates to the technical field of automatic detection of cancer cells, in particular to a method for identifying cancer cells based on an improved U-net convolutional neural network model. Background technique [0002] With the rapid development of computer science and technology, people's requirements for computers are constantly improving. At present, they are not only satisfied with massive data storage, search and massive data calculation, but people begin to hope to use computers to realize some specific skills of our human beings. , so the field of computer application continues to increase. Among them, the use of computer technology for medical image processing and analysis and automatic identification has played an extremely important role in medical diagnosis. [0003] When analyzing medical images, it is one of the most basic methods for diagnosing diseases to use microscopes to observe manually and diagnose the shape and numbe...

AI Technical Summary

Problems solved by technology

[0003] When analyzing medical images, it is one of the most basic methods for diagnosing diseases to use microscopes to observe manually and diagnose the shape and number of red blood cells, white blood cells, etc., but manual inspections like this Blood micrographs are a time-consuming and energy-intensive endeavor

At the same time, in the traditional image processing method, the method of processing the image according to some operators is limited by many conditions, resulting in the inability to accurately and efficiently identify cancer cells

Images