We performed multi-resolution analysis using the wavelet transform on components of surface reflections from a facial image to identify wrinkles and fine asperities. Additionally, by applying principal component analysis, we statistically analyzed relationships between distribu- tions of the wrinkles and surface asperities, and actual ages. In previous research, the components of facial surface reflection were calculated as for multi-resolution analysis, however, the uneven illuminance on the face affected the trend. Therefore, we proposed a method to transform the luminance unevenness of facial images into a uniform distribution using signal processing, which contributes to the appropriate analysis of the surface reflection components on the face. As a result, compared with previous research, we could analyze the lighting unevenness with less influence of bias and acquire the distribution tendency of wrinkles and fine asperities in each direction.
View full abstract