2d image to 3d model using Machine Learning (Automated)
Add information (initially default values) to given 2D image to make 3D model that exist in given image.
Fact: In rendering processing many information about 3D model is used to produced 2D image, but this information not included in 2D output. So 2D to 3D process we have to add many information have to add back to 3D model. Idea : Machine Learning process can help us to make such information to build 3D model.
2 Comments
These data sets can be huge, not sure we should add them to the default blender. Also I do see benefits being this kind of technologies outside blender so should perhaps be developed outside, but using an add-on to interface with it.
This solution would also need a render farm in order to create those data sets, before this kind of solution can work. Perhaps some parties would be interested in delivering such render-farm. It doesn't seem fair that Blender should be investing in those render farms.
Its a great idea, but it would make blender huge with all those libraries