To exactly perform intelligent encoding/decoding to an image to be encoded corresponding to the expression/motion of a model image.
Concerning the method for performing intelligent encoding and decoding to images A1, A2... in which a subject having an expression change exists, at the time of encoding, a 1st image A1, in which a 1st subject having a reference expression exists, and a 2nd image B1, in which a 2nd subject having a non-reference expression exists, are prepared, a deformation parameter is generated for deforming the 2nd subject in the 2nd image B1 from the reference expression to the non-reference expression and the 1st image A1 and the deformation parameter are encoded. At the time of decoding, the 1st subject in the 1st image A1 is deformed from the reference expression to the non- reference expression while using the deformation parameter.
HARASHIMA HIROSHI
KANEKO MASAHIDE
SAITO TAKAHIRO
TOSHIBA CORP