Moscow University Automates Mismatch Detection

By Tom Allen
subscribe

The Graphics & Media Lab CMC of Lomonosov Moscow State University has developed ‘a practical algorithm for automatic channel mismatch detection in S3D video’. The algorithm was used to evaluate more than 100 S3D films in ‘a reasonable amount of time’. 22% of tested content contained at least one scene with channel mismatch; the results show that such scenes can appear even in high-budget projects. Detailed information will be published later this year.

Definition:

Channel mismatch is an uncommon but important issue for S3D films; it can be defined as a stereoscopic image with swapped left and right views, either fully or partially, causing some or all objects to have incorrect depth values. While the human brain compensates for this, it leads to mental fatigue and visual discomfort.