Implementation of a TCM-based computational health informatics diagnostic tool for Sub-Saharan African students

Abstract

Abstract Health status checkup is a crucial step towards early detection of diseases. Health status diagnosis, in university health centers, within the sub-Saharan African region, can be cumbersome and time consuming. In many cases, facilities for health checkup are not available. Traditional Chinese Medicine (TCM) is a promising approach, when integrated with in-silico methods. This study was conducted to implement a TCM-based computational health informatics diagnostic tool. The tool was applied to diagnose African students. This study was also conducted to stimulate further research into in-silico TCM diagnostics. Besides developing a reliable biometric verification system, to ascertain the real identities of patients brought to university health centers, it is assistive to create a platform that provides automated and complementary support for preliminary health diagnostic activities. It also mitigates stress, by helping to efficiently decipher and provide quick objective opinion from the perspective of a computerized decision support system. The diagnostic module of the computational health informatics diagnostic tool adopts knowledge from a TCM facial color diagnosis. A comprehensive literature search was conducted for relevant full-text research papers. Only research publications written in English language were reviewed. The present work was compared qualitatively and quantitatively with the existing works noted in the literature. Facial detection and matching algorithms were implemented for the TCM-based computational health informatics diagnostic tool by using Java programming language. Facial image acquisition processes were conducted. Captured facial images of African students were preprocessed. Facial feature extraction was performed by implementing feature extraction algorithms. An algorithm for the extraction of color information and measurement was also implemented. Knowledge of machine learning was applied to extract and collate facial features, and to machine learn from them. Facial classification and recognition algorithms were implemented. Finally, the results from the computational health informatics diagnostic tool were evaluated, by conducting a performance evaluation and validation. This study provides qualitative and quantitative information on facial recognition, facial color information measurement, as well as prediction of health status, for some sub-Saharan African University students. Performance evaluation was shown using confusion matrix and ROC curves. Statistical analysis of the experimental results was presented. The parameters in each diagnostic illustration were shown with valid range. In order to justify the effectiveness of the computational tool, further explanations were provided from relevant methodology guides on the evaluation of diagnostic tests. The computational health informatics diagnostic tool will complement the diagnostic efforts in university health centers of sub-Saharan African universities. It will also be useful for personal health diagnosis of interested individuals. The tool will also be viable for educating health professionals. TCM will be of immense benefit to developing countries by positively contributing towards diagnosing different non-communicable diseases and some infectious diseases in such countries.