[Project] [Paper]

If you use this work, please cite:

@inproceedings{ren-wacv2017,
    title = {Who moved my cheese? Automatic Annotation of Rodent Behaviors with Convolutional Neural Networks},
    author = {Ren, Zhongzheng and Noronha, Adriana and Ciernia, Annie Vogel and Lee, Yong Jae},
    booktitle = {Winter Conference on Applications of Computer Vision (WACV)},
    year = {2017}
}

• Linux
• NVIDIA GPU + CUDA CuDNN Recommended(CPU mode and CUDA without CuDNN also work)

• Install [Caffe]
• Install [C3D]
• Install [pythonVideoAnnotator]
• Clone this repo

Please visit this [google drive folder]

You can use the UI in this code to annotate your own videos. It can be used to annotate any videos with dense frame-level action labels. We also provide the tools to parse the annotations and generate txt files for caffe to use.

We fine-tune an alexnet and C3d model here. Basically, change the softmax to 5 class and fine-tune the whole networks.

If you meet any trouble when using this program, fell free to contact me!