VHDL Realization of Efficient H. 264 Intra Prediction Scheme based on Best Prediction Matrix Mode

Abstract

This paper introduces a VHDL realization of a new efficient intra prediction scheme that aims to enhance the compression efficiency of the H.264 standard. The new proposed algorithm is called Best Prediction Matrix Mode (BPMM). The main idea behind the new prediction scheme is to combine the most usable intra prediction modes, {vertical -horizontal -DC}, into a new efficient prediction mode. The performance of the new proposed prediction scheme with respect to compression ratio, Peak Signal to Noise Ratio (PSNR) and bit rate is evaluated. The results show that the BPMM enhances the compression ratio and correspondingly the bit rate and it noticeably increases the PSNR.

International Journal of Electronics and Telecommunications

“…The key daunting trouble is that we have a high overhead, a result of the existence from 240 to 960 extra bits in each subblock mode would provoke a heavy loss of bandwidth. Accordingly, the decreasing overhead is a bottleneck for (BPMM) in practice [21].…”

“…It should be noted that the best prediction matrix is an excellent substitute of the standard prediction modes and we will rely on it only to achieve the prediction. Further details regarding the proposed algorithm can be found in [9,10]. As shown in Table II the preliminary findings have shown that the (BPMM) had achieved a leading compression ratio with an average of 33.56%.…”

“…Fairly recently, the authors proposed the architectural framework for Best Prediction Matrix Mode (BPMM) [9]. This mode of prediction has relied on an established fact that each of DC, vertical and horizontal modes are most used than the others.…”

“…More details on (BPMM) are covered in the next section. Our proposed technique had gone through a series of three main phases as follows: phase I the proposal stage in which the BPMM achieved better compression ratios, bit rates, and PSNR depending on improving the Intra prediction scheme of H.264 and reducing the residual information by selecting best values for each pixel [9,10].…”

Performance evaluation of a new efficient H.264 intraprediction scheme

“…The absolute difference module is responsible for calculating the absolute difference values. Finally, the comparator module finds the minimum value of the absolute differences and then selects a value between (upper, left, upper-left corner, and average) that corresponds to this minimum and stores it in the best prediction matrix [24].…”