| 摘要 |
PURPOSE: To enable correspondent allocation for reducing the degree of human intervention for morphing, matching and object recognition. CONSTITUTION: A feature point 22 is discriminated from a source point 20 and a target point, the mode form of a total kinetic mode at the body part points of respective elastic bodies simulated equally with the position of the feature point is calculated for each source image and each target image (28, 30 and 32), a feature vector expressing the relation of body parts to the feature point is related with the respective feature points, and the target feature points related to the generalized feature vectors are respectively allocated to plural source feature points (40) and interpolated between the position and value of the source feature point and the target point so that an intermediate image can be generated (44). |
