舌诊是中医望诊的重要手段,同时,温度与人体的健康息息相关。为了研究舌面的脏腑功能定位及舌象温度关系的反映,论文提出了一种红外技术的感兴趣区域(region of interest,ROI)模型研究方法。首先,利用葛立恒扫描法和Bezier曲线对多边形ROI模型进行改进;然后,借助U-Net分割网络将提取出的温度信息进行训练与学习,从而做到批量处理舌体温度信息;最后,利用HSV色彩模型进行3D可视化,达成舌象温度分区的可视化研究。此外,为了验证该方法的准确性,实验还对模型截取出的舌体进行了评价指标验证,准确度可以达到0。991 1,分割效果极佳。研究表明:改进后的红外信息提取技术既能直观地观察到舌体的分区状况,也可以完整保留舌体的信息变化,为中医的数据化提供了完整可行性方案。实现了舌体红外信息数据的提取与中医诊断技术的有机结合。解决了中医一体化望诊的舌体信息完整性及准确性问题。
Research for an Improved ROI Model of Tongue Infrared Information
Tongue diagnosis is an important tool for Chinese medicine inspection,while temperature is closely related to human health.We propose a region of interest(ROI)model research method for infrared technology based on the functional localization of the internal organs on the tongue surface and the reflection of the tongue temperature relationship.Firstly,the polygonal ROI model is improved by using Graham's scan method and Bezier curve.Then,the extracted temperature information is trained and learned with the help of U-Net segmentation network,so as to achieve batch processing of tongue temperature information.Finally,the HSV color model is used for 3D visualization to reach the visualization study of tongue temperature partitioning.In addition,in order to verify the accuracy of the proposed method,the experiment also verified the evaluation index of the tongue body truncated by the model,and the accuracy could reach 0.991 1,with excellent segmentation effect.The study shows that the improved infrared information extraction technique can not only visualize the partitioning condition of the tongue body,but also retain the information changes of the tongue body intact,which provides a complete feasible solution for the datafication of Traditional Chinese Medicine.The organic combination of extraction of infrared information data of the tongue body and Traditional Chinese Medicine diagnosis technology is realized,which solves the problem of completeness and accuracy of the tongue information in the integrated Chinese medicine inspection.
tongue temperatureGraham's scanBezier curveROI modelU-Net network modelvisualizationtongue information