Classification of CT scan and X-ray dataset based on deep learning and particle swarm optimization

• Published in: PloS one Vol. 20; no. 1; p. e0317450
• Main Authors: Liu, Honghua, Zhao, Mingwei, She, Chang, Peng, Han, Liu, Mailan, Li, Bo
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 27.01.2025; Public Library of Science (PLoS)
• Subjects: Accuracy; Algorithms; Artificial intelligence; Betacoronavirus; Biology and Life Sciences; Classification; Computed tomography; Computer and Information Sciences; Computer-aided medical diagnosis; Coronavirus Infections - diagnostic imaging; Coronaviruses; COVID-19; COVID-19 - diagnostic imaging; CT imaging; Datasets; Deep Learning; Disease detection; Disease transmission; Early warning systems; False Positive Reactions; Genetic algorithms; Humans; Image processing; Image Processing, Computer-Assisted - methods; Lung - diagnostic imaging; Lung diseases; Lungs; Machine learning; Medical imaging; Medical research; Medicine and Health Sciences; Methods; Monitoring; Network management systems; Neural networks; Objective function; Optimization algorithms; Optimization models; Pandemics; Particle Swarm Optimization; Physical Sciences; Pneumonia, Viral - diagnostic imaging; Public health; Radiology; Research and Analysis Methods; Research methodology; ROC Curve; SARS-CoV-2; Telemedicine; Tomography, X-Ray Computed - methods; X-rays; China
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0317450

In 2019, the novel coronavirus swept the world, exposing the monitoring and early warning problems of the medical system. Computer-aided diagnosis models based on deep learning have good universality and can well alleviate these problems. However, traditional image processing methods may lead to high false positive rates, which is unacceptable in disease monitoring and early warning. This paper proposes a low false positive rate disease detection method based on COVID-19 lung images and establishes a two-stage optimization model. In the first stage, the model is trained using classical gradient descent, and relevant features are extracted; in the second stage, an objective function that minimizes the false positive rate is constructed to obtain a network model with high accuracy and low false positive rate. Therefore, the proposed method has the potential to effectively classify medical images. The proposed model was verified using a public COVID-19 radiology dataset and a public COVID-19 lung CT scan dataset. The results show that the model has made significant progress, with the false positive rate reduced to 11.3% and 7.5%, and the area under the ROC curve increased to 92.8% and 97.01%.