Multi-domain medical image translation generation for lung image classification based on generative adversarial networks

• Published in: Computer methods and programs in biomedicine 2023-02, Vol.229, p.107200-107200, Article 107200
• Main Authors: Chen, Yunfeng, Lin, Yalan, Xu, Xiaodie, Ding, Jinzhen, Li, Chuzhao, Zeng, Yiming, Xie, Weifang, Huang, Jianlong
• Format: Article
• Language: English
• Subjects: Algorithms; Area Under Curve; COVID-19 - diagnostic imaging; Humans; Image processing; Image Processing, Computer-Assisted; Key transfer branch; Lung - diagnostic imaging; Lung classification; MI-GAN; Multi-domain translation; Pneumonia, Viral
• ISSN: 0169-2607; 1872-7565
• DOI: 10.1016/j.cmpb.2022.107200

•Medical image multi-domain translation model MI-GAN based on the focus transfer generation branch is proposed.•Conversion between domains ensures the attention performance of the medical image multi-domain translation model and the quality of the synthesized images.•Pneumonia classification network based on InceptionV3 pretraining model, training data includes real images and synthetic images.•Model can effectively improve the efficiency and accuracy of image classification. Lung image classification-assisted diagnosis has a large application market. Aiming at the problems of poor attention to existing translation models, the insufficient ability of key transfer and generation, insufficient quality of generated images, and lack of detailed features, this paper conducts research on lung medical image translation and lung image classification based on generative adversarial networks. This paper proposes a medical image multi-domain translation algorithm MI-GAN based on the key migration branch. After the actual analysis of the imbalanced medical image data, the key target domain images are selected, the key migration branch is established, and a single generator is used to complete the medical image multi-domain translation. The conversion between domains ensures the attention performance of the medical image multi-domain translation model and the quality of the synthesized images. At the same time, a lung image classification model based on synthetic image data augmentation is proposed. The synthetic lung CT medical images and the original real medical images are used as the training set together to study the performance of the auxiliary diagnosis model in the classification of normal healthy subjects, and also of the mild and severe COVID-19 patients. Based on the chest CT image dataset, MI-GAN has completed the mutual conversion and generation of normal lung images without disease, viral pneumonia and Mild COVID-19 images. The synthetic images GAN-test and GAN-train indicators reached, respectively 92.188% and 85.069%, compared with other generative models in terms of authenticity and diversity, there is a considerable improvement. The accuracy rate of pneumonia diagnosis of the lung image classification model is 93.85%, which is 3.1% higher than that of the diagnosis model trained only with real images; the sensitivity of disease diagnosis is 96.69%, a relative improvement of 7.1%. 1%, the specificity was 89.70%; the area under the ROC curve (AUC) incre