search

autonomousvision / LaRa

[ECCV 2024] Efficient Large-Baseline Radiance Fields, a feed-forward 2DGS model
https://apchenstu.github.io/LaRa/
MIT License
277 stars 12 forks source link

Input of the method #6

Closed chenhonghua closed 3 months ago

chenhonghua commented 3 months ago

Hi,

Are the results generated by mvdream suitable for this approach? If so, how can mvdream be adapted to this framework?

apchenstu commented 3 months ago

Thanks for the question. Our method is a feed-forward multi-view reconstruction method, so the input is basically posed multi-view images. The multi-view images can be real captured images or generated images. You can also generate images by mvdream.

chenhonghua commented 3 months ago

Thanks!