Abstract:

To optimize the intraprediction of the H.264 coding standard for improving the speed of intraprediction,a 4×4 block pipeline is adopted instead of serial processing method.A adaptive pipeline is adopted which harmonizes the intraprediction under different blocks and different prediction modes,using the fourchannel parallel multifunction processing element to calculate the prediction values of four pixels in a column in one cycle.The “seed” method is adopted to replace the way of calculate directly in planar prediction mode.The optimized intraprediction circuit is designed by using FPGA development platform.The verification results show that the optimized intraprediction circuit only needs 196 clock cycles in average to process a macro block.

Key words: H 264;pipeline;intraprediction;multifunction processing element