| 摘要 |
PURPOSE: A method for matching a face image using a panorama image in a model-based encoding system is to match a front image, both sides images, and a rear image of a user's face with a three dimensional standard face image. CONSTITUTION: A matching method comprises the steps of: overlapping a front image, both sides images and a rear image of a user's face with a text face image corresponding to the front image, the both sides images and the rear image; converting a standard face image by a movement displacement and a predesignated scale factor; matching the standard face image with the text face image; detecting positions of characteristic elements such as eyes, nose, mouth and ears from the text face image and extended or contracted three-dimensional standard face image; and matching the characteristic elements within the text face image with corresponding characteristic elements within the extended or contracted three-dimensional standard face image. |
