2004•IEICE Transactions on CommunicationsRequires access

Fast reference frame selection method for motion estimation in JVT/H.264

Ching-Ting Hsu, Hung-Ju Li, Mei‐Juan Chen

Open publisher page 6 citations

Abstract

The three main reasons why the new H.264 (MPEG-4 AVC) video coding standard has a significant performance better than the other standards are the adoption of variable block sizes, multiple reference frames, and the consideration of rate distortion optimization within the codec. However, these features incur a considerable increase in encoder complexity. As for the multiple reference frames motion estimation, the increased computation is in proportion to the number of searched reference frames. In this paper, a fast multi-frame selection method is proposed for H.264 video coding. The proposed scheme can efficiently determine the best reference frame from the allowed five reference frames. Simulation results show that the speed of the proposed method is over two times faster than that of the original scheme adopted in JVT reference software JM73 while keeping the similar video quality and bit-rate.

About this research paper

What this paper is about

The three main reasons why the new H.264 (MPEG-4 AVC) video coding standard has a significant performance better than the other standards are the adoption of variable block sizes, multiple reference frames, and the consideration of rate distortion optimization within the codec. However, these features incur a considerable increase in encoder complexity. As for the multiple reference frames motion estimation, the increased computation is in proportion to the number of searched reference frames. In this paper, a fast multi-frame selection method is proposed for H.264 video coding. The proposed scheme can efficiently determine the best reference frame from the allowed five reference frames. Simulation results show that the speed of the proposed method is over two times faster than that of the original scheme adopted in JVT reference software JM73 while keeping the similar video quality and bit-rate.

Why it matters

OpenAlex reports 6 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The three main reasons why the new H.264 (MPEG-4 AVC) video coding standard has a significant performance better than the other standards are the adoption of variable block sizes, multiple reference frames, and the consideration of rate distortion optimization within the codec. However, these features incur a considerable increase in encoder complexity. As for the multiple reference frames motion estimation, the increased computation is in proportion to the number of searched reference frames. In this paper, a fast multi-frame selection method is proposed for H.264 video coding. The proposed scheme can efficiently determine the best reference frame from the allowed five reference frames. Simulation results show that the speed of the proposed method is over two times faster than that of the original scheme adopted in JVT reference software JM73 while keeping the similar video quality and bit-rate.

Key concepts: Computer science, Reference frame, Codec, Reference software, Coding (social sciences), Encoder, Residual frame, Block-matching algorithm

Related papers

Back to paper searchBrowse research topicsOriginal source
Fast reference frame selection method for motion estimation in JVT/H.264 — Research Paper | ScholarLens