Vast Amounts of Video Data Clean Algorithm Base on Bradley-Terry Model

DOI

10.2991/emcm-16.2017.21How to use a DOI?

Keywords

Bradley-terry model; ELO rating system; Video quality assessment; Data cleaning

Abstract

Traditional video data cleaning method is to use video quality subjective labeling to achieve. This method relies on the subjective and objective factors such as the level of expert knowledge of the observer and the observing environment, so that a high accuracy rate cannot be ensured, and in the massive data conditions need to consume a lot of manpower and time. In this paper, Bradley-Terry model is applied to the field of video cleaning for the first time, let each observer will randomly selected two videos from the sample than Choose the better of the two according to their quality, then we use the ELO Rating System points the results, finally an ordered set is obtained. We use the real-time monitoring of video data to complete the experiment and evaluate our algorithm with stability and effectiveness. The results show that the proposed video quality sorting method has good stability and accuracy, and can be finely and quickly sort the data by their quality in the video big data environment. Experimental results show that the algorithm can clean the video data accurately and quickly under the condition of massive video.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Binwen Cao
AU  - Li Chen
AU  - Hui Lie
PY  - 2017/02
DA  - 2017/02
TI  - Vast Amounts of Video Data Clean Algorithm Base on Bradley-Terry Model
BT  - Proceedings of the 2016 7th International Conference on Education, Management, Computer and Medicine (EMCM 2016)
PB  - Atlantis Press
SP  - 104
EP  - 110
SN  - 2352-538X
UR  - https://doi.org/10.2991/emcm-16.2017.21
DO  - 10.2991/emcm-16.2017.21
ID  - Cao2017/02
ER  -