In-situ implementation of adaptive color constancy algorithm

Color constancy is a concept used in various computer vision applications, such as image segmentation, object recognition, scene classification, and image forensics. This work speaks about designing an adaptive color constancy algorithm which identifies the regions of skin found in faces. Illumination level is estimated by training and is corrected using proposed algorithm with different source of illumination, thus proving the variation in skin color pixels under different illumination conditions. This algorithm is implemented on images captured using webcam under different illuminations. It is helpful to estimate the illumination and correct the same with respect to the reference image under natural light.