|
| 1 | +# An Adaptive Screen-Space Meshing Approach for Normal Integration |
| 2 | +### [Project Page](https://moritzheep.github.io/adaptive-screen-meshing/) | [Paper](https://arxiv.org/abs/2409.16907) | [Poster](https://moritzheep.github.io/publication/heep-isotropicmeshing-2024/poster.pdf) | [Video](https://www.youtube.com/watch?v=6m2SKqb1M5M) |
| 3 | + |
| 4 | + |
| 5 | +[Moritz Heep](https://moritzheep.github.io/), |
| 6 | +[Eduard Zell](http://www.eduardzell.com/) |
| 7 | + |
| 8 | +University of Bonn, PhenoRob |
| 9 | + |
| 10 | +# Getting Started |
| 11 | +To clone the repository with all its submodules run |
| 12 | +```Shell |
| 13 | +$ git clone --recursive https://github.com/moritzheep/adaptive-screen-meshing.git |
| 14 | +``` |
| 15 | + |
| 16 | +## Prerequisites |
| 17 | +Our method uses [nvdiffrast](https://github.com/NVlabs/nvdiffrast) to translate between the triangle mesh and the pixel grid. Please make sure that all dependencies of nvdiffrast are met, especially torch. Furthermore, we require [OpenCV](https://opencv.org/) to be installed. |
| 18 | +### Docker |
| 19 | +We prepared a Docker image to take care of these dependencies and facilitate testing. We still need the nvidia-container-runtime. It can be installed via |
| 20 | +```Shell |
| 21 | +$ cd docker |
| 22 | +$ .\nvidia-container-runtime.sh |
| 23 | +``` |
| 24 | +To build the image, run |
| 25 | +```Shell |
| 26 | +$ cd docker |
| 27 | +$ ./build.sh |
| 28 | +``` |
| 29 | +Finally, run |
| 30 | +```Shell |
| 31 | +$ docker run \ |
| 32 | + --runtime=nvidia \ |
| 33 | + --gpus all \ |
| 34 | + adaptive-screen-meshing |
| 35 | +``` |
| 36 | +to get a list of options. You can then mount a volume and point towards your input files. All arguments can be appended to the above command and are passed through. |
| 37 | +## Building |
| 38 | +To build the project, run |
| 39 | +```Shell |
| 40 | +$ mkdir build |
| 41 | +$ cd build |
| 42 | +$ cmake .. |
| 43 | +$ make |
| 44 | +``` |
| 45 | + |
| 46 | +## Running |
| 47 | +After the build has been completed successfully, you can run |
| 48 | +```sh |
| 49 | +$ src/main \ |
| 50 | + -n <path-to-the-normal-map> \ |
| 51 | + -m <path-to-the-foreground-mask> \ |
| 52 | + -t <path-to-save-the-mesh> |
| 53 | +``` |
| 54 | +for a quick test. The normal map should be in the `.exr` format, the mask can be any 8Bit grayscale format supported by OpenCV. The mesh can be saved to any format supported by the [pmp-library](https://github.com/pmp-library/pmp-library). We recommend `.obj`. |
| 55 | + |
| 56 | +You can run `src/main` to get of full list of all options. |
| 57 | + |
| 58 | +# Troubleshooting |
| 59 | +If you get meshes that curve in the wrong direction, try flipping the x or y coordinate of your normal map. |
| 60 | + |
| 61 | +# Citation |
| 62 | +``` |
| 63 | +@inproceedings{heep2024screen-space-meshing, |
| 64 | + title={An Adaptive Screen-Space Meshing Approach for Normal Integration}, |
| 65 | + author={Moritz Heep and Eduard Zell}, |
| 66 | + booktitle = {European Conference on Computer Vision (ECCV)}, |
| 67 | + year={2024} |
| 68 | +} |
| 69 | +``` |
0 commit comments