An approach towards a full-reference-based benchmarking for quality-optimized endoscopic video stabilization systems
Marvin C. Offiah, Navya Amin, Thomas Groß, Nail El-Sourani, Markus Borschbach · 2012
Since the last two decades, research in the area of video stabilization has been tremendously increasing. More and more algorithms are being developed for counteracting the distortions existing in the videos. In the field of invasive diagnostics and therapy, as a result of manual handling of the diagnostic apparatus, occurrence of these distortions is unavoidable leaving the resulting videos speculative for examiners. Therefore, development of stabilization algorithms, specifically for the endoscopic videos, has become crucial. Despite of the stabilization process, the resultant videos are left with some distortions, leaving the stabilization algorithm unimpressive. As yet reduced-reference-based metrics for stabilized video quality assessment have been developed. Objective assessment of these videos using full-reference videos could help accurately benchmark the stabilization algorithms which this paper focuses on. A comparison of stabilized versions of synthetically distorted videos to their synthetic full-reference videos is performed in the current study.