2008Unpublished venueRequires access

Fast multiple reference frame selection algorithm in H.264/AVC

Changwoo Ha, Wan-Jae Lee, Joohyun Lee, Jechang Jeong

Open publisher page 0 citations

Abstract

The H.264 video coding standard provides considerable performance improvement in the adoption of variable block sizes, multiple reference frames, and the consideration of rate distortion optimization within the codec compared to previous standards. However, these features increase the complexity of motion estimation in the encoder. As for multiple reference frames motion estimation, the increased computation is in proportion to the number of searched reference frames. In this paper, we focused on a complexity reduction algorithm in multiple reference frame motion estimation (MRF_ME). A fast MRF_ME algorithm determines a criterion to quickly terminate the search processing for multiple reference frames. The experimental results showed that the proposed method improved the motion estimation time over the implementation of the H.264 reference software while maintaining a similar PSNR and bit-rate.

About this research paper

What this paper is about

The H.264 video coding standard provides considerable performance improvement in the adoption of variable block sizes, multiple reference frames, and the consideration of rate distortion optimization within the codec compared to previous standards. However, these features increase the complexity of motion estimation in the encoder. As for multiple reference frames motion estimation, the increased computation is in proportion to the number of searched reference frames. In this paper, we focused on a complexity reduction algorithm in multiple reference frame motion estimation (MRF_ME). A fast MRF_ME algorithm determines a criterion to quickly terminate the search processing for multiple reference frames. The experimental results showed that the proposed method improved the motion estimation time over the implementation of the H.264 reference software while maintaining a similar PSNR and bit-rate.

Why it matters

A significance statement is not available in the OpenAlex record.

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 H.264 video coding standard provides considerable performance improvement in the adoption of variable block sizes, multiple reference frames, and the consideration of rate distortion optimization within the codec compared to previous standards. However, these features increase the complexity of motion estimation in the encoder. As for multiple reference frames motion estimation, the increased computation is in proportion to the number of searched reference frames. In this paper, we focused on a complexity reduction algorithm in multiple reference frame motion estimation (MRF_ME). A fast MRF_ME algorithm determines a criterion to quickly terminate the search processing for multiple reference frames. The experimental results showed that the proposed method improved the motion estimation time over the implementation of the H.264 reference software while maintaining a similar PSNR and bit-rate.

Key concepts: Reference frame, Motion estimation, Reference software, Inter frame, Rate–distortion optimization, Codec, Computer science, Quarter-pixel motion

Related papers

Back to paper searchBrowse research topicsOriginal source
Fast multiple reference frame selection algorithm in H.264/AVC — Research Paper | ScholarLens