An interactive virtual mirror to support makeup for visually impaired persons

Because visually impaired persons are not able to confirm the appearance of their own face, they are afraid of and uneasy about makeup. We have been developing a system that assists makeup application through verbal feedback according to the appearance of the user's face. The system encourages social communication by helping the user feel confident. In this paper, we introduce a new method of using a dedicated camera device and an image processing algorithm to quantify lip makeup. We designed the camera device to capture face images under different illumination types; white light and green light. In the image processing, we extract the lip area from a segmentation that uses the difference in saturation between two lighting conditions. We also developed a symmetry histogram-based geometrical feature of lip shape to estimate the symmetrical characteristics. The experimental results show that the proposed approach yields results close to a human's subjective evaluation.