YouTube UGC Dataset for Video Compression Research
This addresses the lack of public UGC datasets for video compression research, which is crucial for improving quality assessment in video sharing applications, though it is incremental as it focuses on dataset creation and metric discussion.
The paper tackles the problem of inaccurate video compression and quality assessment for User Generated Content (UGC) by introducing a large-scale dataset of 1500 video clips from YouTube, covering categories like Gaming and Sports, and demonstrates that no-reference metrics like Noise, Banding, and SLEEQ offer a promising evaluation method.
Non-professional video, commonly known as User Generated Content (UGC) has become very popular in today's video sharing applications. However, traditional metrics used in compression and quality assessment, like BD-Rate and PSNR, are designed for pristine originals. Thus, their accuracy drops significantly when being applied on non-pristine originals (the majority of UGC). Understanding difficulties for compression and quality assessment in the scenario of UGC is important, but there are few public UGC datasets available for research. This paper introduces a large scale UGC dataset (1500 20 sec video clips) sampled from millions of YouTube videos. The dataset covers popular categories like Gaming, Sports, and new features like High Dynamic Range (HDR). Besides a novel sampling method based on features extracted from encoding, challenges for UGC compression and quality evaluation are also discussed. Shortcomings of traditional reference-based metrics on UGC are addressed. We demonstrate a promising way to evaluate UGC quality by no-reference objective quality metrics, and evaluate the current dataset with three no-reference metrics (Noise, Banding, and SLEEQ).