Deep learning-based occupational pneumoconiosis multi-modal analysis method

Abstract

The invention provides a deep learning-based occupational pneumoconiosis multi-modal analysis method, and belongs to the field of pneumoconiosis analysis, and the method comprises the steps: collecting chest X-film image information and personal basic information of a person; performing word vectorization processing on the personal basic information; constructing a one-dimensional convolutional neural network and a two-dimensional convolutional neural network, and establishing a multi-modal convolutional neural network (MM-CNN) model on this basis; taking the two kinds of information as input of the MM-CNN model, establishing a multi-classification MM-CNN pneumoconiosis analysis model, and forming an objective function when corresponding constraints are met; optimizing hyper-parameters of the multi-classification MM-CNN pneumoconiosis analysis model by using a shuffled frog leaping algorithm (SFLA); using the optimized multi-classification MM-CNN pneumoconiosis analysis model to analyze the chest X-film image information of the person and the personal information subjected to word vectorization processing, and outputting an analysis result. According to the method, accurate and real-time detection and analysis of lung health of people can be realized, and early warning of part of occupational pneumoconiosis is completed.

Description

technical field [0001] The invention belongs to the field of pneumoconiosis analysis, in particular to a multimodal analysis method for occupational pneumoconiosis based on deep learning. Background technique [0002] In recent years, my country's production safety has achieved a sustained and stable improvement, and the number of deaths from production safety accidents has dropped rapidly for many years in a row. However, the situation facing occupational health work is still very severe. As a kind of occupational disease, occupational pneumoconiosis is mainly distributed in industrial industries such as coal, nonferrous metals, machinery, building materials, and light industry. For example, in the process of coal mine production, a lot of dust (mainly including coal dust and silica dust) will be generated in many links such as rock roadway blasting, rock roadway loading, rock roadway excavation, coal roadway blasting, coal roadway reinforcement, coal preparation and transp...

AI Technical Summary

Problems solved by technology

Although the CNN model has excellent performance, how to design a suitable network structure and find the best model parameters for specific application problems is a difficult problem in the application process of CNN

Images