Depth continuity estimation in S3D video
- Author: Alexander Bokov
- Supervisor: dr. Dmitriy Vatolin
One of the features that contribute to a movie’s visual quality is the arrangement of objects in the scene. The most important (salient) objects are usually placed near the center of the screen, and sudden changes of the region of visual attention are avoided to make the watching experience more comfortable. In stereoscopic videos, the same requirement applies to the depth field. We have developed a metric for estimating the smoothness of transitions in disparity during scene cuts.
Here is an illustration of the depth continuity metric. For each frame of the video we plot the parallax range, expressed in percent of frame width. Brighter areas on the chart correspond to more common disparity values in the shot. The top example shows well-aligned shots, the depth jump cut will not cause any discomfort. The bottom example contains a noticeable depth jump cut that could have become better upon more careful alignment.
The developed metric aims to estimate the depth jumps while assuming a simple visual saliency model based on two cues: center prior and defocus. The final value produced by the metric for a movie is dimensionless; higher values indicate more frequent and intense depth jump cuts throughout the movie.
We computed the overall depth continuity metric for 105 movies (detailed analysis will soon be available in report 10 under all reports). The following diagram depicts the metric value for each movie relative to its release date:
MSU Benchmark Collection
- MSU Video Upscalers Benchmark 2022
- MSU HDR Video Reconstruction Benchmark 2022
- MSU Super-Resolution for Video Compression Benchmark 2022
- MSU Video Quality Metrics Benchmark 2022
- MSU Video Alignment and Retrieval Benchmark
- MSU Mobile Video Codecs Benchmark 2021
- MSU Video Super-Resolution Benchmark
- MSU Shot Boundary Detection Benchmark 2020
- MSU Deinterlacer Benchmark
- The VideoMatting Project
- Video Completion
- Codecs Comparisons & Optimization
- Video Quality Measurement Tool 3D
MSU Datasets Collection
- Video Filters