Deep learning-based stereo camera multi-video synchronization

TL;DR

This paper explores deep learning methods for software-based synchronization of stereo camera videos, aiming to replace hardware solutions with a more flexible, cost-effective approach.

Contribution

It compares various deep learning models for stereo video synchronization and demonstrates their potential for practical, production-ready software solutions.

Findings

Some models are efficient and generalizable for synchronization

Deep learning can effectively replace hardware synchronization methods

The study advances towards practical software-based stereo video synchronization

Abstract

Stereo vision is essential for many applications. Currently, the synchronization of the streams coming from two cameras is done using mostly hardware. A software-based synchronization method would reduce the cost, weight and size of the entire system and allow for more flexibility when building such systems. With this goal in mind, we present here a comparison of different deep learning-based systems and prove that some are efficient and generalizable enough for such a task. This study paves the way to a production ready software-based video synchronization system.