The system processes video footage to detect violent activities in real-time. It uses a combination of Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN - LSTM) to classify video sequences and compute a mean violence score for each sequence
Steps:
Step 1: Move Files
Navigate to the lstm_model directory and run the 1_move_file.py script to move all files into the appropriate train/test folders.
Step 2: Extract Files
Run the 2_extract_files.py script to extract images from the videos and create a data file for training and testing.
Step 3: Extract Features
Navigate to the data directory inside the lstm_model folder and run extract_features.py to generate extracted features for each video.
Step 4: Train Model
In the same data directory, run train.py to train the LSTM model. The trained model will be saved in the checkpoints directory.
Step 5: Run Web Application
Navigate to the root directory of the project and run main.py to start the web application: python main.py
Ensure the saved_model path variable in main.py points to the correct location of the trained model
.png?versionId=zGBI0HCIzD952w.E3LdHIsKx2vJmrLq3&Expires=1795407402&Key-Pair-Id=K3EVG0QTJR4SK0&Signature=tg3YnKpnr~K~j5VpWtl7JUPk1f1nwgRW~QTpArhck40TiwdQwDlBxyQAbYBJeLv0eLpXbJKVG0GVgsub3sXMhQXgmOvUyJWZmtbgR6iB9FW8kf7sjfkzZn~dZKk4-zACAEF6DdfiRO2P8PJMQdAONO3QWEpuEmaVSH9njZzVhN7~fDAR1BpmtONu4MiFJKRRIbgoNFVntLPM-4QCPSKML0702Qe-TNrg~sW4URh8ZPLHsMT36ZP1WGHx2V4bxMGiwhAiOcIceMU3UN6t1656d3KcVdA0aC4nq0YNUWJMzXr6Ex9KL-VFb78BT6AmJjkBO8TKLkRe0uknuMgoNJz8EA__)