Abstract
Hairstyle is one of the most important features people use to characterize one's appearance. Since whether a hairstyle is suitable or not is relied on the aesthetic sense of human, mathematical model could not represent the suitability. We focused on the fact that the suitable hairstyle is related to one's shape. This paper proposes a new technique for automatically retrieving a suitable hairstyle from a collection of hairstyle examples through learning the relationship between facial shapes and suitable hairstyles. A method of hair-face image composition utilizing matting technique was also developed to synthesize realistic hairstyle images. A database depending on users' preference was constructed to retrieve more suitable hairstyles. The effectiveness of the proposed technique was validated through evaluation experiments by themselves and the others.
