RISA-Net: Rotation-Invariant and Structure-Aware Network for Fine-grained 3D Shape Retrieval

Rao FU, Jie Yang, Jiawei Sun, Fanglue Zhang, Yu-Kun Lai and Lin Gao. Project Page

1) Fine-grained 3D shape retrieval dataset

Please download our fine-grained 3D shape retrieval dataset. The dataset provides a quantatitive measure for fine-grained 3D shape retrievals. It contains 6 object categories: knife, guitar, car, plane, chair and table, each of which is further divided into dozens of categories. We provide 5 versions of the datset:

1. unregistered integerated aligned model,
2. unregistered integerated perturbed model,
3. unregistered segmented aligned model,
4. unregistered segmented perturbed model,
5. regitered segmented aligned model.

2) Training

We provide code to train the RISA-Net.

a) Preprocessing

We first need to extract the base geometric feature: edge length and diheral angles from the registed segmented shapes. We also need to analyse structure information and make a lable file for triplet loss training. All preprocessing codes are placed in the matlab file. Please install Matlab before running the code.

• To extract base geometric feature, please run: get_edge_feature_all.m.
• To analyse structure information, please run: pca_of_each_part.m.
• To make label file for triplet loss, please run: gmake_label_for_trip,m.

b) Learning

Our network is based on Tensorflow. First, you need to set up an environment. Please run:

cd python;
pip install -r requirements.txt

After the environment is set up, you can train our network. Please run:

python risanet.py -a 1e3 -b 1e2 -c 1e0 -d 1e3 -e 1e2 -f 5000 -x 0.3 -y 0.3 -s 32 -m 32 -n 32

After the network is trained, you can load the shape descriptors for shape retrieval. Please run:

python risanet.py -r /path/to/checkpoint -k 5000

3) Demos

Here we provide some retrieval results on several datasets.

Citation

If you use this code for your research, please consider citing:

hahaha