Label Distribution Learning Forest

Publication

Code accompanying the paper Label Distribution Learning Forest. [arXiv] [ProjectPage]

If you use our code, please consider citing following paper:

@inproceedings{shen2017label,
  title={Label Distribution Learning Forests},
  author={Shen, Wei and Zhao, Kai and Guo, Yilu and Yuille, Alan},
  booktitle={Advances in neural information processing systems},
  year={2017}
}

How to use

Clone source and build

git clone https://github.com/zeakey/LDLForest --recursive
cd LDLForest/caffe-ldl
cp Makefile.config.example Makefile.config
make -j$(nproc) && make pycaffe && cd ..

Please customize your own Makefile.config before make.

Try out LDL toy data:

1. Download LDL datasets:

   bash data/ldl/get_ldl.sh

2. Try out LDL example:

   python experiments/ldl/run.py

Facial age estimation:

1. Download the Morph dataset

2. Make LMDB database for Morph:

   python tools/morph2lmdb.py --data /path/to/Morph

3. Start to train:

   python experiments/alexnet/run.py

A live demo.

Visit the project page for pretrained models and more detailed information.