The five video classification methods:

1. Classify one frame at a time with a ConvNet
2. Extract features from each frame with a ConvNet, passing the sequence to an RNN, in a separate network
3. Use a time-dstirbuted ConvNet, passing the features to an RNN, much like #2 but all in one network (this is the lrcn network in the code).
4. Extract features from each frame with a ConvNet and pass the sequence to an MLP
5. Use a 3D convolutional network (has two versions of 3d conv to choose from)

See the accompanying blog post for full details: https://medium.com/@harvitronix/five-video-classification-methods-implemented-in-keras-and-tensorflow-99cad29cc0b5

1. This code requires you have Keras 2 and TensorFlow 1 or greater installed. Please see the requirements.txt file. To ensure you're up to date, run:

pip install -r requirements.txt

2. You must also have ffmpeg installed in order to extract the video files. If ffmpeg isn't in your system path (ie. which ffmpeg doesn't return its path, or you're on an OS other than *nix), you'll need to update the path to ffmpeg in data/2_extract_files.py.

3. Configuration of the runs is performed using yottato, from

https://github.com/prabindh/yottato

After cloning or download, perform below steps to all to local python package list.

cd yottato

python setup.py install

Important configuration parameters {location of data, hyperparameters} are configurable via the JSON file at,

config/config.json

Typically, the below parameters would need to be configured. Note - the instructions in below section (Getting the data, or running different commands) apply to the already set default parameters.

• globalDataRepo : This locates a central place where media files are stored, and where results of analysis are kept

DEFAULT Repo location: ./data

• training/algorithm : This identifies the model/algorithm to be used (ex, lrcn, lstm, cnn etc)

DEFAULT algorithm : lrcn

First, download the dataset from UCF into the data folder:

cd data && wget http://crcv.ucf.edu/data/UCF101/UCF101.rar

Then extract it with unrar e UCF101.rar.

Next, create folders (still in the data folder) with mkdir train && mkdir test && mkdir sequences && mkdir checkpoints.

Now you can run the scripts in the data folder to move the videos to the appropriate place, extract their frames and make the CSV file the rest of the code references. You need to run these in order. Example:

python 1_move_files.py

python 2_extract_files.py

Before you can run the lstm and mlp, you need to extract features from the images with the CNN. This is done by running extract_features.py. On my Dell with a GeFore 960m GPU, this takes about 8 hours. If you want to limit to just the first N classes, you can set that option in the file.

The CNN-only method (method #1 in the blog post) is run from train_cnn.py.

The rest of the models are run from train.py. Configurations can be performed using the config/config.json

The models are all defined in models.py. Reference that file to see which models you are able to run in train.py.

Training logs are saved to CSV and also to TensorBoard files. To see progress while training, run tensorboard --logdir=data/logs from the project root folder.

• Add data augmentation to fight overfitting
• Support multiple workers in the data generator for faster training
• Add a demo script
• Support other datasets
• Implement optical flow
• Implement more complex network architectures, like optical flow/CNN fusion

Khurram Soomro, Amir Roshan Zamir and Mubarak Shah, UCF101: A Dataset of 101 Human Action Classes From Videos in The Wild., CRCV-TR-12-01, November, 2012.