Chest X-ray quality evaluation method fusing medical field knowledge

• Main Authors: YE CHENGJIE, FENG RUI, HUANG GUOYING, JIANG LONGQUAN, SHEN QUANLI, ZHANG XIAOBO, XU HONG, WANG YINGWEN, FU WEIJIA, HUANG XINYU, GE XIAOLING
• Format: Patent
• Language: Chinese; English
• Published: 09.06.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATIONTECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING ORPROCESSING OF MEDICAL OR HEALTHCARE DATA; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS

The invention provides a chest X-ray quality evaluation method fusing medical domain knowledge, which comprises the following steps of: converting text data of a chest X-ray diagnosis text report and a quality control report into triple information through knowledge extraction and knowledge fusion, and taking the triple information as domain knowledge guidance information corresponding to a medical image to participate in model training. According to the method, large-scale medical images, diagnostic reports and quality control report text labeling data are utilized, and cross-domain transfer learning is realized by performing fine adjustment on a pre-training model based on a comparison text-image pair, so that the problem of insufficient training sample data volume is effectively solved; and block local visual features of the X-ray image are aligned with a plurality of text features, so that the visual features are ensured to contain more fine-grained image information, and the problem that global visual in