Seiko Epson Corporation

A.i. Soft, Inc.

Kazuya Horii

A method and apparatus are provided for merging plural facial images by continuously varying the shape interpolation ratio while preventing the positions of facial features from shifting, retaining a specific resemblance between the synthesized image and either of the source images in specific facial features, and changing the facial expression using a simple operation. The facial image processing apparatus comprises first shape matching processor 11 for grossly matching a facial image G1 with a background facial image G3 by focusing on the feature areas of facial image G1 and converting the coordinates of particular points in the feature areas in the shape model of this facial image G1 to the coordinates of the corresponding point in the shape model of the background facial image G3; a second shape matching processor 12 for grossly matching a facial image G2 with a background facial image G3 by focusing on the feature areas of facial image G2 and converting the coordinates of particular points in the feature areas in the shape model of this facial image G2 to the coordinates of the corresponding point in the shape model of the background facial image G3; shape interpolation processor 1 for interpolating the coordinates of corresponding points in the matched shape models of facial images G1 and G2 and the shape model of background facial image G3 at a specific interpolation ratio; and color interpolation processor 2 for interpolating the color of each pixel in the shape model interpolated by shape interpolation processor 1 using a specific interpolation ratio.

Growing community of inventors

Suwa, Japan

Kazuya Horii