DeepMind recently released a new paper called “Neural Scene Representation and Rendering”. Its pretty amazing stuff, I have to admit. Their AI, if given just a single image of an environment, can generate a 3D map of that environment. That includes all the objects, textures, colors, and shapes that it sees. In this video,. I’ll talk about its applications, theory, and code it up myself in Keras even though they haven’t released any code to give you a programmatic example.
Code for this video:
Please Subscribe! And like. And comment. That’s what keeps me going.
Join us in the Wizards Slack channel:
Sign up for the next course at The School of AI:
And please support me on Patreon: