AUTOMATIC SEGMENTATION OF HAIR IN IMAGES

摘要

Based on a multi-step process, an automatic hair segmentation method and system are disclosed. By extracting various information components from an image, including background color, face position, hair color, skin color, and skin mask, a heuristic-based method is created for the detection and segmentation of hair that can detect hair with an accuracy of approximately 75% and with a false hair overestimation error of 34%. Furthermore, it is shown that down sampling the image down to a face width of 25 px results in a 73% reduction in computation time with insignificant change in detection accuracy.

主权项

1. A computer-implemented method to detect hair pixels in a digital image, the method comprising:
detecting, by at least one processor, a face location, face rotation and facial features in the digital image; loading, by the at least one processor, a hair prior likelihood mask and aligning the hair prior likelihood mask to the face location and face rotation; computing, by the at least one processor, a hair color probability mask based on colors of expected hair regions around the face location, and combining, by the at least one processor, the hair prior likelihood mask and the hair color probability mask, to compute a final hair likelihood mask.