Action-Based Multi-Camera Synchronization

We propose a video alignment method based on observing the actions of a set of articulated objects. Given object association information, the proposed video synchronization method is applicable to general and unconstrained scenarios in a way that is not feasible with current state-of-the-art approaches: the proposed method does not impose constraints on the relative pose or motion of the cameras, on the structure of the time warping between the videos and on the amount of overlap among the fields-of-view. The proposed method uses a high-level video analysis (object actions) and models the alignment as a frame association problem (as opposed to the traditional continuous time warping). We present a qualitative and quantitative analysis of the results in real-world complex scenarios, showing the robustness of the method and higher accuracy compared to the only approach from the literature that works under similar conditions.