This repository contains the code for the CVPR 2020 paper "Differentiable Volumetric Rendering: Learning Implicit 3D Representations without 3D Supervision"
hello, I want to try you approach with my own data, I have rgb and calibrated depth camera, could you give me some advice on how to create data like your dtu dataset, thanks very much.
hello, I want to try you approach with my own data, I have rgb and calibrated depth camera, could you give me some advice on how to create data like your dtu dataset, thanks very much.