Tongue diagnosis is an important method in both Traditional Chinese Medicine (TCM) and Western Medicine (WM), as the tongue's appearance can reflect a person’s overall health. Among the key features observed, tongue shape and color play a major role in identifying certain diseases and tracking their progression. This study focuses on the tongue image analysis method of artificial intelligence (AI) to detect shapes and colors of tongue for fast health screening without any need for human intervention. The proposed system firstly uses a deep learning YOLOv10 (You Only Look Once) on 750 tongue images in four tasks. The first task uses the YOLOv10 model to detect and isolate the entire tongue region from the input image ensure that the following tasks focus only on the tongue region. Second, to accurately classify the tongue into seven shape categories, including normal, geographic, fissured, scalloped, thin, swollen, and deviated tongues. Third, to detect the crack type associated with fissured tongue, including side crack, vertical crack, deep crack and irregular crack. Last, to detect whether the tongue contains ulcers or spots or not. The study also uses CatBoost machine learning algorithm to train 5550 color images captured at different color saturations and under different light conditions and classified into seven classes (red, yellow, green, blue, gray, white, and pink) using several color space models, including (RGB, YcbCr, HSV, LAB, and YIQ) to analyze and extract tongue color features. The Web-based application was developed using Streamlit to offer an easy-to-use graphical interface and provides an automatic tongue shape and color detection tool and compares results based on both TCM and WM perspectives. This study is to support early screening and medical analysis in a fast and reliable way.
Building similarity graph...
Analyzing shared references across papers
Loading...
Ali Al‐Naji
Javaan Chahl
Electrical engineering technical journal.
Building similarity graph...
Analyzing shared references across papers
Loading...
Al‐Naji et al. (Mon,) studied this question.
www.synapsesocial.com/papers/68a36f840a429f79733322d5 — DOI: https://doi.org/10.51173/eetj.v2i2.31
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: