본문 바로가기
  • Home

The Development of Realtime Smart Image Mixing System Using GPU Based Video Codec

  • Journal of Knowledge Information Technology and Systems
  • Abbr : JKITS
  • 2016, 11(6), pp.617-624
  • Publisher : Korea Knowledge Information Technology Society
  • Research Area : Interdisciplinary Studies > Interdisciplinary Research
  • Published : December 31, 2016

Seung-Taek Ryoo 1 Kim, Youung Kwon 2 Chang, Jae-Khun 1

1한신대학교
2오산대학교

Accredited

ABSTRACT

As a result of recent advances in GPU programming using graphics hardware, many studies have been focused on the application of parallel and realtime rendering. This paper introduces the development of realtime smart image mixing system using ultra high density video based decoding and encoding techniques. The smart image mixing system implements sequential and streaming based image mixer that consists of video decoding, image resizing and mixing, video encoding step. Sequential based image mixing system decodes whole frames from multiple video inputs and merges extracted synchronous frames from multiple videos into single image. And then, encodes each single image into whole frames and save final result video file. Streaming based image mixing system decodes single frame from multiple video inputs and merges an extracted synchronous frame from multiple videos into single image. And then, encodes a merged image into each frame and repeat the above steps(decoding/mixing/encoding) until last frame from multiple video inputs is extracted. We compare sequential based image mixing method with streaming based image mixing method. In temporal performance, streaming based image mixing method is better than sequential based image mixing method. Suggested method can be used in various fields such as image and game contents generation, UHD image display and network service and video conferencing.

Citation status

* References for papers published after 2023 are currently being built.