Questions about cases on dynamic video NVS

Question

Questions about cases on dynamic video NVS

Closed this issue 4 months ago · 8 comments

xizaoqu commented 4 months ago

Hi, thanks for sharing the great work.

When testing, I find that some cases will result in weird quality, do you have some insights about this?

warp_video.mp4

diffusion2.mp4

Answer 1 · 2024-09-14T04:16:39.000Z

Hi, our model is designed to handle static images rather than dynamic videos. We will release a model for video NVS soon; please stay tuned.

Answer 2 · 2024-09-14T06:26:08.000Z

@xizaoqu
Hi, I would like to ask what do you use as prompt to generate this video, is it a single image with text as prompt？

Answer 3 · 2024-09-14T06:28:50.000Z

@Drexubery
Hello author, what does video NVS mean? What are the input and output respectively? When will it be released? Can you give me an approximate time?

Answer 4 · 2024-09-15T00:38:27.000Z

Hi, our model is designed to handle static images rather than dynamic videos. We will release a model for video NVS soon; please stay tuned.

Thanks.

Answer 5 · 2024-09-15T00:39:12.000Z

you use as prompt to generate this video, is it a single image with text as promp

Hi, it is a video from the DAVIS dataset.

Answer 6 · 2024-09-15T02:06:51.000Z

@xizaoqu Thank you for your reply! Sorry to bother you again. Do you mean that the input is a video？(I'm curious because the viewcrafer currently only supports a single image as input.)

Answer 7 · 2024-09-15T02:10:51.000Z

@xizaoqu Thank you for your reply! Sorry to bother you again. Do you mean that the input is a video？(I'm curious because the viewcrafer currently only supports a single image as input.)

The origin input is an image, but it will process it into a video (you can check the output file/rendering.mp4).

Answer 8 · 2024-09-15T02:17:15.000Z

@xizaoqu Thanks again!