New coronal pneumonia severe prediction method and system based on machine learning algorithm

Abstract

The invention belongs to the technical field of machine learning, and particularly relates to a new coronal pneumonia severe prediction method based on a machine learning algorithm, which comprises the following steps: acquiring a lung CT image of a patient, and segmenting and extracting a lung effective region; constructing a RetinaNet convolutional neural network model, training the model, inputting the lung effective region into the model, performing regression on the severe degree, and outputting a Pct value; collecting clinical structural data of a patient, carrying out dimension reduction processing on the data, inputting the data into the trained random forest model, carrying out prediction, and outputting a Pclinical value; and fusing the Pct value and the Pclinical value through Fbeta calculation to obtain the final severe degree of the patient. The defects in the prior art are overcome, a machine learning algorithm is used for prediction, and the severity degree of the new coronal pneumonia is predicted by using images and structural data through machine learning.

Description

technical field [0001] The invention belongs to the technical field of machine learning, and in particular relates to a method and system for predicting the severity of new coronary pneumonia based on machine learning algorithms. Background technique [0002] COVID-19 has spread rapidly around the world, causing serious loss of life and property around the world. In the detection and treatment of COVID-19, in order to reduce the mortality rate, the prediction of patients with severe pneumonia is of great significance. During the treatment of new coronary pneumonia, CT images play an important role in the diagnosis of new coronary pneumonia. Patients take CT images regularly to observe the development and changes of the disease. [0003] The Chinese invention with the patent application number 202010492779.9 discloses a prediction algorithm for the severity of new coronary pneumonia based on 25(OH)D levels. The degree of prediction, this method only uses some clinical medica...

AI Technical Summary

Problems solved by technology

[0004] The Chinese invention with patent application number 202010269407.X discloses an assessment method based on quantitative ultrasound analysis, which reduces the professional and technical requirements for doctors to perform ultrasound. This method only uses the B-line changes of ultrasound images to score the severity of the disease , its actual auxiliary diagnosis effect needs to be verified

Images