REGION-BASED PROCESSING OF PREDICTED PIXELS

摘要

A method for decoding a compressed video data sequence containing one or more coded pixel blocks. The compressed video sequence is buffered. Prediction information for each of the coded pixel blocks is reviewed. One or more groups of coded pixel blocks are formed based on the reviewed prediction information such that the coded pixel blocks within a given group have similar prediction dependencies and/or at least do not depend on a reconstructed pixel within a group of received pixel blocks to enable parallel decoding. The formed groups are scheduled for processing and subsequently decoded to produce a decoded video data sequence.

主权项

1. A decoding method for coded video data, comprising:
assigning a first coded pixel block of a frame to be decoded to a decoding group; for subsequent additional coded pixel blocks of the frame;
determining whether the respective subsequent coded pixel block contains a prediction reference that refers to a reconstructed pixel within the decoding group,if not, assigning the respective subsequent coded pixel block to the decoding group, andif so, terminating the decoding group; retrieving from a memory previously decoded video data referenced by the prediction references of the decoding group; and decoding the coded pixel blocks of the decoding group with to the retrieved decoded video data.