zouchuhang / 3d-prnn Goto Github PK

View Code? Open in Web Editor NEW

68.0 7.0 19.0 135.64 MB

Torch implementation of our ICCV 17 paper: "3D-PRNN, Generating Shape Primitives with Recurrent Neural Networks"

Home Page: http://openaccess.thecvf.com/content_ICCV_2017/papers/Zou_3D-PRNN_Generating_Shape_ICCV_2017_paper.pdf

License: MIT License

Lua 31.46% MATLAB 68.49% M 0.04%

shape-generation 3d-prnn deep-learning

3d-prnn's Introduction

3D-PRNN

Torch implementation of our ICCV 17 paper: "3D-PRNN, Generating Shape Primitives with Recurrent Neural Networks"

Prerequisites

Linux
NVIDIA GPU + CUDA CuDNN
Torch

matio: https://github.com/tbeu/matio

distributions: https://github.com/deepmind/torch-distributions
Matlab (for visualization)

Data

Download primitive data to current folder

This includes our ground truth primitives (folder "prim_gt") and the original ModelNet mesh (folder "ModelNet10_mesh")

Train

For shape generation from scratch:

th driver.lua

For shape generation conditioned on single depth map:

th driver_depth.lua

Generation

For shape generation from scratch:

th testNet_3dp.lua

For shape generation conditioned on single depth map:

th testNet_3dp_depth.lua

Visualization

To visualize ground truth primitives, run visualizeGTPrimitive.m in Matlab
To visualize sample shape generation, run visualizeRandomGeneration.m
To visualize sample shape generation conditioned on depth, run visualizeDepthReconGeneration.m

Primitive ground truth

See ./matlab/ folder

Note

For shape generation conditioned on depth, as explained in the paper Sec 5.1, we perform a nearest neighbor query based on the encoded feature of the depth map to retrieve the most similar shape in the training set and use the configuration as the initial state for the RNN. For convenience, we include our pre-computed initial configuration for each test class in folder "data/sample_generation".

Primitive parsing

We provide in the matlab folder the demo code (demo.m) to parse single primitive. To sequentially parse primitives in batch, see "script_parse_primitive.m". After each run of "script_parse_primitive.m", run "script_parse_primitive_symmetry.m" to get the symmetry. With every three parses, try "script_refine_parse_primitive.m" to refine the parsed primitives.

Citation

@inproceedings{zou20173d,
  title={3d-prnn: Generating shape primitives with recurrent neural networks},
  author={Zou, Chuhang and Yumer, Ersin and Yang, Jimei and Ceylan, Duygu and Hoiem, Derek},
  booktitle={The IEEE International Conference on Computer Vision (ICCV)},
  year={2017}
}

Acknowledgement

We express gratitudes to the torch implementation of hand writting digits generation as we benefit from the code.

3d-prnn's People

Contributors

Stargazers

Watchers

Forkers

johndpope deeprrl jwzxgy2007 claudehang shubhampachori12110095 mannykayy songyankuirobo-mind hippogriff tony-hou peterzs hailieqh nnu-gisa aymanmukh dritchie mobmonrob phymucs aoe-khkhan eurekamoments

3d-prnn's Issues

Code for Primitive fitting.

Can you please provide code for primitive fitting?

Can you explain the data format?

Hi, I want to re-trained the network on my own data and I notice that each array entry in prim_gt folder is of shape (num_of_parts, 20). Can you explain what do these 20 parameters mean?

Also, what's the use for prim_rnn_batch_nz_all_tr.t7 and prim_rnn_batch_nz_all_val.t7?

Thanks in advance!

Error in primitive fitting code

Thank you for releasing the code. There are few error that I am getting while using the script script_parse_primitive.m.

The VoxelPlotter function is not there. However, I found out the code on the internet. Similarly there is a missing source file in minFunc_2012. That too I found online.
Loading previously parsed primitives is not working (that is re-running the script script_parse_primitive.m with previously parsed primitive), because of the following reasons:
a. In the script script_parse_primitive, there is error "Reference to non-existent field 'prim_all" for line 29.
b. In the same script at line 71, "theta_prev = prim_param(vox_num_q,20);", there is an error about exceeding the dimension. I guess this is because it is trying to access 20th index, which is absent.

It seems that the pipeline is broken in between different script, e.g. script_parse_primitive.m, script_refine_parse_primitive.m and script_parse_primitive_symmetry.m. If you can provide solution to the second point, that will be great.

cannot get data_3dp.zip through wget

When I use wget http://czou4.web.engr.illinois.edu/data/data_3dp.zip to download the data, I can only get data_3dp.zip of size 8kb. It seems the zip file is broken.
Are there any other choices to download the data, such as google drive?

What's the recommended environment?

Hi, I'm trying to run the code but met some problems which I believe is due to the environment. Could you please provide the environment you are using? Like the version of Ubuntu, CUDA, torch and other required packages.