Automatic local color correction in S3D video

Introduction

When shooting stereoscopic video, many factors can cause color mismatch between camera views, such as illuminated camera filters, glare, polarized light, etc. We have developed an automatic method for elimination of color mismatch between stereo video views. A comparison with analogues showed that it has a higher-quality result and works faster.

Color mismatch in released "Spy Kids 3D: Game Over", 0:14:19

Example of color distortion, "Spy Kids 3D: Game Over", 0:14:19


Color difference between views

Color difference between views


Result of automatic correction by the proposed algorithm

Corrected colors

Proposed method

The proposed approach includes:

Experiments

The algorithm was compared to Ocula 3.0 (for Nuke 7.0) in Minimum Correction and Average Correction modes, and YuvSoft Stereo Processing Suite Pro 1.0 (Adobe After Effects CS 5.5)

Objective comparison

Ocula 3.0 Average Ocula 3.0 Minimum YuvSoft SPS 1.0 Proposed method
SSIM 0.9981 0.9980 0.9951 0.9992
Y-PSNR 31.786 39.772 33.289 45.707
Working time, FPS 0.09 0.11 0.25 1.37

Artificially made distortions

Before color correction

Artificial distortion


Result of the proposed algorithm

Automatically corrected colors


Comparison between algorithms (difference with original frame)

Comparison of color difference

Distortions in real video

Frame from "Pirates of the Caribbean: On Stranger Tides" trailer

Distortions in real video


Result of the proposed algorithm

Automatically corrected colors


Comparison between algorithms (difference with compensated frame)

Comparison of color difference

29 May 2019
See Also
Call for HEVC codecs 2019
Fourteen modern video codec comparison
Parallax range estimation in S3D video
Parallax determines the depth of S3D movies. The range of parallaxes should be both comfortable and entertaining for spectators.
Geometric distortions analysis and correction
Production of low-budget movies is prone to errors. Our method automatically corrects rotation and scale mismatch.
Automatic detection of artifacts in converted S3D videos
Our set of algorithms detects edge sharpness mismatch, cardboard effect, and crosstalk noticeability.
Temporal shift estimation for stereoscopic videos
How to take into account geometric distortions in the estimation of the temporal shift?
Neural network-based algorithm for classification of stereoscopic video by the production method
What method was used to create the 3D scene?
Site structure