Awesome
Deep Repulsive Clustering of Ordered Data Based on Order-Identity Decomposition
Official TensorFlow Implementation of the ICLR 2021 paper, "Deep Repulsive Clustering of Ordered Data Based on Order-Identity Decomposition."
Requirements
- TensorFlow 2.0 or higher
- python 3.7
Pretrained Model and Reference List for Quick Start
Datasets
- For MORPH II experiments, we follow the same fold settings in this OL repo.
Quick Start: Code Usage Example
- Modify Config file
- Adjust img_folder, train & test list for your purpose. As a default, MORPH setting A is used in the source code.
- Clustering ordered data by DRC-ORID
$ cd train
$ cd morph
$ python train_kCH_morph_clustering.py
- This will generate the centroids file and checkpoint of feature extractor.
- Get clustering results and train VGG-based network
$ python get_clustering_info.py
$ python train_kCH_morph_estimation.py
- Select references based on ORID results
$ cd test
$ python morph_ref_sel_kCH_by_attr.py
- Run test
$ python test_morph_kCH_by_attr.py
Cite
@inproceedings{lee2021repulsive},
title = {Deep Repulsive Clustering of Ordered Data Based on Order-Identity Decomposition},
author = {Lee, Seon-Ho and Kim, Chang-Su},
booktitle = {International Conference on Learning Representations},
year = {2021}
}
License
See Apache License