Tongue analysis holds promise for disease detection and health monitoring,especially in traditional Chinese medicine.However,its subjectivity hinders clinical applications.Deep learning offers a path for automated ton...Tongue analysis holds promise for disease detection and health monitoring,especially in traditional Chinese medicine.However,its subjectivity hinders clinical applications.Deep learning offers a path for automated tongue diagnosis,yet existing methods struggle to capture subtle details,and the lack of large datasets hampers the development of robust and generalizable models.To address these challenges,we introduce TonguExpert(https://www.biosino.org/TonguExpert),a free platform for archiving,analyzing,and extracting phenotypes from tongue images.Our deep learning framework integrates cutting-edge techniques for tongue segmentation and phenotype extraction.TonguExpert analyzes a massive dataset of 5992 tongue images from a Chinese population and extracts 773 phenotypes including five predicted labels and their probabilities,355 global features(entire tongue,tongue body,and tongue coating)and 408 local features(fissures and tooth marks)in a unified process.Besides,580 additional features for five tongue subregions are also available for future study.Notably,TonguExpert outperforms manual classification methods,achieving high accuracy(ROC-AUC 0.89-0.99 for color,0.97 for fissures,0.88 for tooth marks).Additionally,the model generalizes well to predict new phenotypes(e.g.,greasy coating)using external datasets.This allows the model to learn from a broader spectrum of data,potentially improving its overall performance.We also release the largest publicly available dataset of tongue images and phenotypes,which is invaluable for advancing automated analysis and clinical applications of tongue diagnosis.In summary,this research advances automated tongue diagnosis,paving the way for wider clinical adoption and potentially expanding the applications in the future.展开更多
基金the Strategic Priority Research Program of the Chinese Academy of Sciences(Grant no.XDB38020400 to S.Wang and XDB38030100 to G.Zhang)CAS Youth Innovation Promotion Association(2020276 to Q.Peng)+9 种基金CAS Young Team Program for Stable Support of Basic 488 Research(YSBR-077 to S.Wang)Hutian Scholar funded by the Beijing University of Chinese Medicine(2022-XJ-KYQD-003 to J.Chen)CAS Interdisciplinary Innovation Team to S.Wang,Shanghai Municipal Science and Technology Major Project(Grant No.2017SHZDZX01 to S.Wang,L.Jin,J.Wang,J.Tan,and Q.Peng2023SHZDZX02 to L.Jin and G.Zhang)the National Natural Science Foundation of China(NSFC)(32325013,92249302 to S.Wang,32271186 to J.Tan,82004222 to X.Wang,82025036 to J.Chen)the National Key Research and Development Project(2018YFC0910403 to S.Wang)Ministry of Science and Technology of the People’s Republic of China(2015FY111700 to L.Jin)Shanghai Science and Technology Commission Excellent Academic Leaders Program(22XD1424700 to S.Wang)111 Project(B13016)to L.JinCAMS Innovation Fund for Medical Science(2019-I2M-5-066 to J.Wang and L.Jin).
文摘Tongue analysis holds promise for disease detection and health monitoring,especially in traditional Chinese medicine.However,its subjectivity hinders clinical applications.Deep learning offers a path for automated tongue diagnosis,yet existing methods struggle to capture subtle details,and the lack of large datasets hampers the development of robust and generalizable models.To address these challenges,we introduce TonguExpert(https://www.biosino.org/TonguExpert),a free platform for archiving,analyzing,and extracting phenotypes from tongue images.Our deep learning framework integrates cutting-edge techniques for tongue segmentation and phenotype extraction.TonguExpert analyzes a massive dataset of 5992 tongue images from a Chinese population and extracts 773 phenotypes including five predicted labels and their probabilities,355 global features(entire tongue,tongue body,and tongue coating)and 408 local features(fissures and tooth marks)in a unified process.Besides,580 additional features for five tongue subregions are also available for future study.Notably,TonguExpert outperforms manual classification methods,achieving high accuracy(ROC-AUC 0.89-0.99 for color,0.97 for fissures,0.88 for tooth marks).Additionally,the model generalizes well to predict new phenotypes(e.g.,greasy coating)using external datasets.This allows the model to learn from a broader spectrum of data,potentially improving its overall performance.We also release the largest publicly available dataset of tongue images and phenotypes,which is invaluable for advancing automated analysis and clinical applications of tongue diagnosis.In summary,this research advances automated tongue diagnosis,paving the way for wider clinical adoption and potentially expanding the applications in the future.
