摘要 |
PURPOSE: A template matching method and apparatus thereof are provided to scale a space where is used for a template by encoding and decoding images. CONSTITUTION: A current block is selected from a current frame(S610). A first area is selected(S612). A base block is selected(S614). A second area is created by the up-sampling of the selected base block(S616). A template is constituted by combining the first area and the second area(S618). A motion vector is calculated in order to minimize a difference between the template and a part of the frame(S620). A prediction block is determined(S630). The current block is decoded(S640). |