How to Input Multi-View Videos Simultaneously？

I noticed that the video demonstrates the capability to synchronize and process multi-view autonomous driving footage with multi-view fusion. However, I haven’t found a way to input multiple video perspectives at once in the CLI. Could you please guide me on how to ensure multi-view videos can be processed simultaneously?

As shown in the demo in the image below：

Start by inspecting the repository's CLI entry point and the existing video-input handling, then compare those paths with the multi-view fusion demo shown in the issue. Determine whether simultaneous multi-view input is already supported or requires a defined interface; done should mean the supported input procedure is clear and verified, or the requested capability is implemented and tested.

Written by the indexing model from the issue text.