| 摘要 |
The present invention relates to a coding apparatus and method for high-speed processing of HEVC, which determine an optimal coding unit (CU) by using pixel information (a pixel average value, a pixel variance value, and the like) of a main CU block and a sub CU block, and perform prediction only for the determined CU, thereby reducing the number of performances of prediction and thus reducing the amount of calculation for motion estimation. A coding apparatus for high-speed processing of HEVC according to an embodiment of the present invention comprises: a determination unit which determines a CU block at an optimal depth based on pixel information for a CU block at a current depth and pixel information about a plurality of sub-CU blocks, separated from the CU block at the current depth, at a subsequent depth; a prediction unit which performs prediction on a per-prediction unit (PU) basis with respect to the CU block at the optimal depth, and acquires a restoration block for the CU block at the optimal depth; and a conversion unit which converts and encodes a pixel difference value between the restoration block and the CU block at the optimal depth on a per-transform unit (TU) basis. |
