Image segmentation approach for realizing zoomable streaming HEVC video

Yuri Umezaki, Satoshi Goto

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

With the popularity of mobile devices such as mobile phones and cameras, zooming and displaying an arbitrary region of interest (ROI) from a high-resolution video has gained considerable importance. In this study, we implement two methods - partial decoding and tiled encoding - for ROI-based streaming in High Efficiency Video Coding (HEVC). In partial decoding, only the ROI and dependence area are decoded; this approach reduces the calculation cost. In tiled encoding, video frames are split into a grid of tiles, each of which is encoded independently; this approach requires a compatible decoder, and it reduces the streaming bitrate and decoding calculation cost. We evaluated these two methods in terms of the decoding calculation cost, bandwidth efficiency, and video quality. Experimental results show that partial decoding reduced the decoding calculation cost by 35-55%. Tiled encoding showed optimal decoding calculation cost and bandwidth efficiency for a tile size of 128 pixels.

Original languageEnglish
Title of host publicationICICS 2013 - Conference Guide of the 9th International Conference on Information, Communications and Signal Processing
PublisherIEEE Computer Society
DOIs
Publication statusPublished - 2013
Event9th International Conference on Information, Communications and Signal Processing, ICICS 2013 - Tainan
Duration: 2013 Dec 102013 Dec 13

Other

Other9th International Conference on Information, Communications and Signal Processing, ICICS 2013
CityTainan
Period13/12/1013/12/13

Keywords

  • Cropping
  • HEVC
  • Mobile streaming
  • Region of interest
  • Zoomable video

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Signal Processing

Fingerprint

Dive into the research topics of 'Image segmentation approach for realizing zoomable streaming HEVC video'. Together they form a unique fingerprint.

Cite this