目的:对比仪器测量法和图像分析法,分析影响测量皮肤颜色的因素.方法:用测色计和图像分析技术对惰性物体和人体皮肤进行重复性颜色测量,通过测量准确性对比两者的优缺点.通过面部皮肤图像分析,开发一种计算机自动算法,用于图像自动色彩校正.结果:仪器测量法和图像分析法在测量惰性物体颜色时误差较小.测量活体皮肤颜色时,仪器测量法的a*值出现较大的误差.在图像分析法中,间隔时间较长的拍摄也出现较大的误差.对图像的RGB颜色通道进行处理,开发的颜色校正方法可提升26倍的颜色准确性.此颜色校正方法验证了产品12周后的美白功效.结论:图像分析法不接触皮肤,不会产生影响,更适合于活体皮肤颜色的测定.本研究开发的算法提供了一种能大批量自动校正图像色彩的方法,可应用于产品功效验证.%Objective This study was to examine the factors influencing the accuracy of skin color measurement by comparing the instrumental and image analysis methods. Methods Color measurement was conducted repeatedly on an inert object and on human skin in vivo using colorimeter and image analysis techniques, and the accuracy was compared to identify the advantages and disadvantages between the two techniques. A facial imaging system was used to develop a computer algorithm for automatic color correction. Results Both instrumental and image analysis methods resulted in low and similar levels of variability when measuring color from the inert object. The in vivo skin color results, however, showed larger variation in a*values. Large variability was also detected in image analysis method during long-term use. Processing the images in individual RGB color channels, a color correction algorithm was developed which was used to color correct images. A 26-fold reduction in variability was achieved after applying color correction to each image. Application of the color correction methodology to a clinical study had shown a potential to identify the product treatment effect before and after using a skin lightening daily regimen for 12 weeks. Conclusion Image analysis method was shown to be more suitable for in vivo skin color measurement due to its non-skin contact operating conditions. The computer algorithm developed in this study provided an automated method to perform image color correction in large quantities. The technique has the potential for product efficacy substantiation.
展开▼
