Create your own 3D scenes from a set of photos.

Previously, we wrote an article about ADOP, an algorithm that studies a set of photos and turns it into a 3D scene that can be explored. The point cloud rendering is performed by a differentiable renderer using multi-resolution one-pixel point rasterization. The efficient one-pixel point rasterization allows using arbitrary camera models and displaying scenes with well over 100M points in real-time.

Recently, the team has published the algorithm's code so now everyone can create a cool 3D scene from images. ADOP is implemented in C++/CUDA using libTorch. The code comes with compile instructions, that thoroughly explain how to use it, and supplementary materials that are meant to help you to learn more about Approximate Differentiable One-Pixel Point Rendering.

Try ADOP yourself by clicking this link. Also, don't forget to join our new Reddit page, our new Telegram channel, follow us on Instagram and Twitter, where we are sharing breakdowns, the latest news, awesome artworks, and more.