Abstract

Info

Title:: Semantic analysis of lung cancer images based on self-attention generative adversarial network

Author(s):: HU Zhijian; YE Zhengchun; ZHENG Hansen; Office of Digital Coordination and Development Research, Fujian Medical University Union Hospital, Fuzhou 350001, China

Keywords:: lung cancer; self-attention mechanism; generative adversarial network; semantic analysis

Abstract:: Abstract: A self-attention generative adversarial network (SAGAN) is proposed to improve the accuracy of histologicalsubtype prediction for lung cancer cases. After collecting and preprocessing the lung cancer image dataset and dataaugmentation, SAGAN model is trained, where the generator uses self-attention mechanism to strengthen feature extraction,while the discriminator optimizes the generation process. Experimental results show that SAGAN model achieves accuraciesof 0.852 and 0.845 on the training and test sets, respectively, with recall rates of 0.833 and 0.829, outperforming the othermodels. Additionally, the narrow confidence intervals indicate the high stability of SAGAN model in classification. SAGANis helpful for lung cancer image analysis, providing stronger support for clinical decision-making.