How can I do this on Wan Vace?

I know that WAN can be used with pose estimates for textV2V, but I'm not sure whether reference images on videos. The only thing I know that can use the REF image for the video is. A workflow or a resource in Wan Vace would be super helpful!