VFHQ: A High-Quality Dataset and Benchmark for Video Face Super-Resolution
This addresses the need for better training data in video face super-resolution, which is incremental as it focuses on dataset improvement rather than a new method.
The authors tackled the problem of low-quality video face super-resolution (VFSR) by creating VFHQ, a high-quality dataset of over 16,000 clips, and demonstrated that models trained on it produce sharper edges and finer textures compared to those trained on VoxCeleb1.
Most of the existing video face super-resolution (VFSR) methods are trained and evaluated on VoxCeleb1, which is designed specifically for speaker identification and the frames in this dataset are of low quality. As a consequence, the VFSR models trained on this dataset can not output visual-pleasing results. In this paper, we develop an automatic and scalable pipeline to collect a high-quality video face dataset (VFHQ), which contains over $16,000$ high-fidelity clips of diverse interview scenarios. To verify the necessity of VFHQ, we further conduct experiments and demonstrate that VFSR models trained on our VFHQ dataset can generate results with sharper edges and finer textures than those trained on VoxCeleb1. In addition, we show that the temporal information plays a pivotal role in eliminating video consistency issues as well as further improving visual performance. Based on VFHQ, by analyzing the benchmarking study of several state-of-the-art algorithms under bicubic and blind settings. See our project page: https://liangbinxie.github.io/projects/vfhq