Chest X-ray pneumothorax segmentation method based on deep learning

Abstract

The invention provides a chest X-ray pneumothorax segmentation method based on deep learning, and the method comprises the steps: obtaining the image information of chest X-rays, and carrying out thepreprocessing of the image information of the chest X-rays; inputting the preprocessed image information into a deep neural network model, and optimizing the deep neural network model by using a spatially weighted cross entropy loss function; and outputting a segmentation result of the chest X-ray pneumothorax image by utilizing the optimized deep neural network model. According to the method, anend-to-end deep neural network model is trained, and the deep neural network model finds out imaging features depicting pneumothorax through continuous autonomous learning so as to segment out suspected pneumothorax regions. Accurate segmentation of chest X-ray pneumothorax is realized, and accurate pneumothorax segmentation can provide important reference for subsequent treatment of a patient.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a chest X-ray pneumothorax segmentation method based on deep learning. Background technique [0002] Pneumothorax refers to air entering the pleural cavity, causing pneumothorax and lung collapse. It is a high-risk disease among lung diseases. The lung tissue and visceral pleura are usually ruptured due to lung disease or external force, or the tiny emphysema bubbles near the lung surface are ruptured, and the air in the lungs and bronchi escapes into the pleural cavity. If it is not treated in time, once it develops into hemopneumothorax or tension pneumothorax, it will pose a great threat to the patient's life, and the purpose of active treatment measures is to remove the gas in the chest cavity as soon as possible, so that the atrophied lung can be re-expanded as soon as possible, and the lung function can be restored. . [0003] Chest X-ray-based imagin...

AI Technical Summary

Problems solved by technology

Due to the accumulation of a large number of chest X-rays, those patients with critical diseases such as pneumothorax may wait for a long time on the list, thus delaying the treatment

At the same time, long hours of overloaded work prevent doctors from concentrating

[0004] Most of the current research is based on deep learning pneumothorax recognition methods, but there are still blind spots in the pneumothorax segmentation technology for chest X-ray

Images