A python 3 script to build a dense and naively tessellated 3D mesh from a depth image and some functions that might be useful to transform the mesh and read and right off files.
Additionally, it also builds the model's axis aligned bounding box.
Requirements:
- PIL (for naive_mesh.py)
- numpy (for the I/O functions in dataio.py)
Sample image inside "images" directory is from the dataset of the following work:
Sungjoon Choi and Qian-Yi Zhou and Vladlen Koltun. Robust Reconstruction of Indoor Scenes. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015.
usage:
python naive_mesh.py depthimage_filename output_filename