daweidavidwang / GenerativePredictableVoxels

Code release for ECCV16 paper "Learning a Predictable and Generative Vector Representation for Objects"

Testing code release for

R. Girdhar, D. F. Fouhey, M. Rodriguez and A. Gupta
Learning a Predictable and Generative Vector Representation for Objects
In Proc. of European Conference on Computer Vision (ECCV), 2016

@inproceedings{Girdhar16b,
    title = {Learning a Predictable and Generative Vector Representation for Objects},
    author = {Girdhar, R. and Fouhey, D.F. and Rodriguez, M. and Gupta, A.},
    booktitle = {ECCV},
    year = {2016},
}

Pre-requisites

1. Caffe (trained and tested with 97f4536, though should work with the latest version). Clone and install in libs dir.
2. Python libs h5py, matplotlib, mayavi.

Download pre-trained models

Download all the models from dropbox to models/ dir.

Testing using the precomputed networks

$ python src/testing/reconst.py  # stores the prediction in output/ folder

Dataset splits, training code to come soon.