Abstract—Video advertising has become important content for web advertising, especially in mobile devices. However, most mobile devices prohibit an auto-play function for video sequences on mobile web browsers. As a result, video coding standards such as MPEG4 and H.265 cannot be used for video advertising on mobile devices. Therefore, CSS image sprites that can show successive images are widely used to play video sequences for web advertising on mobile devices. This paper proposes a novel rate control scheme to improve the coding efficiency, based on CSS image sprites composed of successive still images coded using the JPEG standard. The proposed scheme controls data allocation to blocks according to the importance of the image regions during rate-distortion optimization of the JPEG encoding process. Subjective evaluation based on the Subjective Assessment Methodology for Video Quality (SAMVIQ) method using datasets created using DAVIS2017 and CG showed that the proposed scheme can reduce the total size of coded data while preserving the subjective image quality.

Index Terms—Subjective video quality, rate control, video advertising.

Risako Aoki, Shun Aoki, Takuro Oki, and Shingo Kobayashi are with Department of Computer Science, Graduate School of Science and Technology, Meiji University, Kawasaki, Japan. Ryusuke Miyamoto is with Department of Computer Science, School of Science and Technology, Meiji University, Kawasaki, Japan. Makoto Iguchi is with CMerTV Corporation, Japan.

Cite:Risako Aoki, Shun Aoki, Takuro Oki, Shingo Kobayashi, Ryusuke Miyamoto, Makoto Iguchi, "A Rate Control Scheme Compatible with a Legacy JPEG Decoder to Reduce Data Size Keeping Subjective Quality for Video Advertising on Mobile Devices," International Journal of Computer and Electrical Engineering vol. 11, no. 3, pp. 133-142, 2019.