3D-model-reconstruction

3D reconstruction from input images.

Webpage is deployed at this (address). (Might be down if we run out of google credits :))

Files:

• Evaluation.ipynb : Evaluation notebook which loads a given model and runs the validation set to get the Average IoU score for the model.
• Load_Matlab.py : Used to visualize the binvox file in Matlab. Not used anymore since we use another script for this currently.
• binvox_rw.py : Used to visualize binvox file generated from the network output and also has methods to manipulate binvox files in python (reference).
• create_dataset.py : Filters data from the entire ShapeNet dataset according to the arguments provided. This uses the metadata in the ShapeNet dataset to filter the models according to the requirements.
• occ_data.py : Main file which consists of the occupancy network, the dataloader and the training loop.
• viewvox : Used for the visulaizion of binvox file. (reference)