ERQA: Edge-Restoration Quality Assessment for Video Super-Resolution
This addresses the need for reliable quality assessment in VSR, which is crucial for applications like surveillance and media, though it is incremental as it builds on existing edge-based approaches.
The paper tackles the problem of assessing detail restoration quality in video super-resolution (VSR) by proposing the ERQA metric, which estimates a model's ability to restore real details based on edge fidelity, with experimental validation on the MSU Video Super-Resolution Benchmark.
Despite the growing popularity of video super-resolution (VSR), there is still no good way to assess the quality of the restored details in upscaled frames. Some SR methods may produce the wrong digit or an entirely different face. Whether a method's results are trustworthy depends on how well it restores truthful details. Image super-resolution can use natural distributions to produce a high-resolution image that is only somewhat similar to the real one. VSR enables exploration of additional information in neighboring frames to restore details from the original scene. The ERQA metric, which we propose in this paper, aims to estimate a model's ability to restore real details using VSR. On the assumption that edges are significant for detail and character recognition, we chose edge fidelity as the foundation for this metric. Experimental validation of our work is based on the MSU Video Super-Resolution Benchmark, which includes the most difficult patterns for detail restoration and verifies the fidelity of details from the original frame. Code for the proposed metric is publicly available at https://github.com/msu-video-group/ERQA.