An innovative algorithm enables searching for videos using image analysis instead of conventional text-based tags.
The search engine, developed by a team from the University of Lincoln, UK, will enable real-time video matching and provide users with more relevant results.
“Everyone uses search engines but currently you are only able to search by text even to search for a video clip, thus some results are far removed from what you were looking for,” said Saddam Bekhet, a PhD student at the School of Computer Science at Lincoln University, explaining the motivation behind the project.
“With the huge volume of data, a smarter video analyser is required to associate semantic tags to the uploaded videos, allowing more efficient indexing and search, including the contents of the video. “
Bekhet and his colleagues Dr Amr Ahmed and Professor Andrew Hunter believe their algorithm, capable of identifying semantic similarities between videos by analysing tiny fractions of video frames, could enhance the search mechanism of YouTube and other online video platforms.
“I want to discover the semantic similarity between videos using the content only,” Bekhet explained. “I adapted some new techniques and found that tiny representative frames could be used to discover similarities. The next stage is to build an effective framework.”
Instead of using full-size video frames, the algorithm takes advantage of data in compressed formats. Not only is such analysis faster, it is also more convenient for online use as the majority of online videos are compressed. Using these features, the algorithm can easily analyse videos in real-time and provide results promptly without the need of complex calculations.
The Lincoln University team has recently introduced a paper named ‘Video Matching Using DC-image and Local Features’. The study has been awarded the Best Student Paper Award at the International Conference of Signal and Image Engineering – a part of the 2013 World Congress on Engineering.