Skip to content

Multi-stream CNN architectures for action detection with actor-centric filtering

License

Notifications You must be signed in to change notification settings

pfabreu/two-stream-networks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Two Stream Attention Convolutional Neural Networks

Contributions

We provide code for Two Stream Action Detection in Keras and each respective extension. Much of the credit of this repo goes to @joaogantunes and @ruipimentelfigueiredo.

Prerequisites

  • python >= 2.7
  • numpy (pip install numpy)
  • ffmpeg built with gstreamer (sudo apt install ffmpeg)
  • OpenCV 2.X or 3.X (build with opencv-contrib and ffmpeg and CUDA support)
  • Pillow (pip install Pillow)
  • CUDA + cuDNN (we used 9 and 7.1, respectively)
  • Tensorflow (for backend GPU) (pip install tensorflow-gpu)
  • Keras >= 2.1.6 (pip install keras)
  • Pandas (pip install pandas) and SciPy (pip install scipy) (because of dependencies)
  • Maplotlib + Seaborn (pip install seaborn) for plots
  • (Optional) To use Keras to draw models you need graphviz and pydot.

Data

  • AVA -- For the AVA dataset due to computational power constraints and for quickly testing the architecture, we made our own split of the dataset called mini-AVA, you can download it here

Models

You can get our Keras models here, includes pre-trained Kinetics, UCF101 models and our AVA models pre-trained on both.

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Releases

No releases published

Packages

No packages published

Languages