#jupyter_notebook #3d #3d_reconstruction #3d_vision #ai #augmented_reality #computer_vision #dataset #deep_learning #machine_learning #neural_network #python #pytorch #tensorflow

https://github.com/google-research-datasets/Objectron

#python #3d_reconstruction #3d_vision #monocular_depth_estimation #monocular_geometry_estimation

MoGe-2 is a powerful tool for estimating 3D geometry from single images. It can create detailed point maps, depth maps, and normal maps with high precision. This model is especially useful because it can predict geometry in metric scale, meaning it provides accurate measurements. It also enhances visual sharpness, making it better than previous versions. Users benefit from MoGe-2 by getting precise 3D information from just one photo, which is helpful for applications like robotics or video games. It's fast and works well with different image sizes.

https://github.com/microsoft/MoGe