PDPC Intra Prediction for Lower-Complexity Video Coding

The patent extracts and removes the Position-Dependent Prediction Combination (PDPC) processing step from the conventional video coding pipeline. By eliminating this specific prediction combination operation, the patent reduces processing complexity while maintaining acceptable video coding performance, thus improving video fluency without significant quality degradation.

The patent applies partial action by selectively removing only the PDPC component from the prediction process rather than eliminating all prediction operations. This partial removal approach maintains sufficient prediction accuracy for video quality while reducing the overall processing complexity and improving video fluency.

The patent converts the potential harm of removing PDPC (which might degrade prediction accuracy) into a benefit by demonstrating that the removal actually improves video fluency with acceptable performance loss. The patent shows that the simplified processing without PDPC achieves a favorable trade-off between complexity reduction and performance maintenance.

The patent changes the processing parameter by disabling the PDPC application mode in the prediction process. This parameter change from enabled to disabled state reduces computational complexity while the patent demonstrates through testing that video coding performance remains within acceptable ranges, thus resolving the contradiction between complexity reduction and performance maintenance.