基于CIELAB的舌诊专用比色卡的设计与验证OA
Design and Validation of a Special Color Comparison Card for Tongue Diagnosis Based on CIELAB
目的 针对舌象图片采集中因设备和光照产生色差的问题,基于CIELAB颜色空间设计舌诊专用比色卡,实现舌象图片的颜色校正,并验证其校正效果.方法 结合真实舌象图片,利用IT8.7/4标准色表设计比色卡,并通过对比实验选取算法进行颜色校正,采用颜色稳定性测试检验色卡的颜色稳定性.通过真实世界图像多光源(3种自然光源、3种人造光源)和多设备(6种智能手机)获取舌象图片1200张,进行主、客观验证:客观验证基于Lab色差值(ΔE)分析校正前后的变化,主观验证对同一受试者,使用舌面仪采集舌象图片作为标准图片,计算不同医师判读校正前、后图像的Kappa系数检验一致性,通过医师对舌色、苔色的判读准确率评估效果.结果 设计了包含17个舌色相关色块的U型舌诊专用比色卡,并通过3×3矩阵和Gamma校正算法实现了颜色校正.客观验证显示,校正后各光源和设备组的舌色色块色差值由5-40下降到0-10(P<0.01),自然光源与白光灯下效果最优,高像素、偏真实还原的手机校正效果更佳;主观验证中,舌色Kappa值由0.19上升至0.44,苔色由0.59上升至0.71,舌色判读准确率从26.73%提升至63.05%,苔色从38.47%提升至70.91%.结论 本研究设计的舌诊专用比色卡可有效减小光照和设备引起的色差,为远程中医诊断、多中心舌象数据库构建及舌象特征客观化分析提供了便捷、准确的解决方案,推动了舌诊客观化研究.
Objective To address the color deviation issues in tongue image acquisition caused by equipment and lighting,a specialized color chart for tongue diagnosis was designed based on the CIELAB color space to achieve color correction of tongue images,and its correction effect was verified.Methods Real tongue images were analyzed to design the color chart in combination with the IT8.7/4 standard color table.The algorithm for color correction was selected through comparative experiments,and the color stability of the color chart was verified via color stability tests.Color correction was performed using a 3×3 matrix and Gamma correction algorithm.A total of 1200 tongue images were obtained under multiple light sources(3 natural light sources and 3 artificial light sources)and with multiple devices(6 types of smartphones)in real-world scenarios for subjective and objective verification.For objective validation,the changes in images before and after correction were analyzed based on the Lab color difference value(ΔE).For subjective validation,tongue images collected by a tongue diagnostic instrument from the same subject were used as the standard reference.The Kappa coefficient was calculated based on the interpretations of pre-correction and post-correction images by different physicians to test the consistency,and the effectiveness was evaluated through the physicians'diagnostic accuracy in judging tongue color and coating color.Results A U-shaped special color comparison card for tongue diagnosis was designed,which included 17 color patches related to tongue colors.The color correction was achieved using a 3×3 matrix and Gamma correction algorithm.Objective verification showed that after correction,the color difference values of tongue color patches in each light source and device group decreased from 5-40 to 0-10 significantly(P<0.01).The effects under natural light and white light were the best,and the correction effect of mobile phones with high pixels and partial real colors was better.In the subjective verification,the Kappa value for tongue color increased from 0.19 to 0.44,and that for coating color rose from 0.59 to 0.71.The accuracy rate of tongue color interpretation increased from 26.73%to 63.05%,and that of tongue coating color increased from 38.47%to 70.91%.Conclusion The specialized color chart for tongue diagnosis effectively reduces color deviations caused by lighting and equipment,providing a convenient and accurate solution for remote TCM diagnosis,construction of multi-center tongue image databases,and objective analysis of tongue image features,thus promoting the objectification of tongue diagnosis.
宋晓慧;桂朝伟;王俊文;郭壮;罗俊超;陈立;马嘉轶;刘孟涵;尹继瑶;王明强;张娟
中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700湖北省武汉市洪山区中医医院 武汉 430070湖北省武汉市洪山区中医医院 武汉 430070中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700中国中医科学院中医基础理论研究所 北京 100700
医药卫生
舌诊比色卡颜色校正智能手机
Tongue diagnosisColor chartColor correctionSmartphone
《世界科学技术-中医药现代化》 2026 (8)
2471-2484,14
国家自然科学基金委员会面上项目(82474393):基于多模态知识图谱构建中医老年体质辨识可解释深度学习模型,负责人:王俊文中医诊疗装备首都卫生发展科研专项(2024-2-4281):和合理疗护膝对膝骨关节炎院外日常治疗的随机对照研究,负责人:王俊文中国中医科学院中医基础理论研究所第十七批基本科研业务费自主选题项目(YZX-202305):马堃教授辨治围绝经期崩漏学术思想及临证经验传承研究,负责人:李敏,王俊文.
评论