Suggestion for improvement.

#11
by RogerReyne - opened

Very good neural network, it would be very nice to have the ability to generate a model from several pictures or photos.

Exactly i was rising this on their Github page,

With the amazing results of Image to 3d generations i can only imagine the amount of accuracy it can bring with multiple reference image support

Ideally a 3d generation should support up to 6 images ( for all angles ) imo

For future if you can guide the 3d generation with ability to specify the view ( like top, front, side etc. ) it would produce more accurate results

PS thanks for this amazing model we love it hopefully it gets the updates it deserve.

Sign up or log in to comment