Dataset for Classification for Tongue Image Tri-Dhat based on Thai Traditional Medicine using Transfer Learning Techniques

Traditional Thai medicine (TTM) is an increasingly popular treatment option. Tongue diagnosis is a highly efficient method for determining overall health, practiced by TTM practitioners. However, the diagnosis naturally varies depending on the practitioner's expertise. In this work, we propose tongue image analysis using raw pixels and artificial intelligence (AI) to support TTM diagnoses. The target classification of Tri-Dhat consists of three classes: Vata, Pitta, and Kapha. We utilize our own organized genuine datasets collected from our university's TTM hospital. Class balance and data augmentation were conducted, and we present analysis approaches and experimental designs. Transfer learning techniques for various pretrained models of Deep Learning were exploited. We used two-tailed paired t-tests and single-factor ANOVA analyzes for performance comparisons. Our work demonstrates that DenseNet121 and Xception models provided the most significant results with cropped image datasets, including DSLR-taken and mobile-taken images. Notably, model ensemble evaluations yielded the highest average predictions, achieving a precision of 0.94, an F1 score of 0.96, accuracy of 0.96, sensitivity of 0.96, and specificity of 0.97, supported by a p-value of 0.0003 from ANOVA analysis. We suggest that our methods could be effectively deployed in real-world scenarios to aid TTM practitioners in their diagnoses.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于泰国传统医学舌像分析的数据集，专注于舌像的Tri-Dhat分类，包括Vata、Pitta和Kapha三个类别。数据集包含从大学医院收集的真实舌像图像，经过类别平衡和数据增强处理，适用于深度学习模型训练，特别是迁移学习技术。实验结果表明，使用DenseNet121和Xception模型在裁剪图像上取得了最佳性能，最高精度达到0.96，为辅助传统医学诊断提供了有效工具。