| 摘要 |
<p>PURPOSE:To use only a proper feature point confirmed by the user for the coding by inputting conforming information after a feature point extracted from a two-dimension front portrait picture and a corrected feature point are confirmed. CONSTITUTION:An image pickup section 11 picks up a picture including a human portrait part and a background part and outputs an input picture signal. An image area extraction section 12 outputs an image area of the human portrait part. A feature point extraction section 13 extracts a feature point. A model mapping section 14 outputs three-dimension model information of the human portrait part. The three-dimension model information is inputted to personal information file 16 and an update section 17. An analysis parameter from an analysis section 15 and the three-dimension information from the personal information file 16 are multiplexed by a multiplexer section 18 and the result is sent to a terminal 19 as a coding output of the picture coder. On the other hand, correction/addition information from a correction/addition information input section 21 is fed to the extraction section 13 and a picture synthesis section 20. Confirming information from an initial input conformation information input section 22 is fed to the extraction section 13, the model mapping section 14 and an operation command section 23. A picture synthesis section 20 applies processing to synthesize the picture signal, the feature point, moving information of the corrected feature point and the information of added feature point onto one screen and outputs the result to a display section 25.</p> |
