CCTV Fight Prediction End-to-End Pipeline

Project Overview

This project aims to detect violence in CCTV camera footage using deep learning and video vision techniques. It uses a dataset of 100 videos each of non-violence and violence, with each video being 3 seconds long.

System Requirements

1. Clone the Git Repository

git clone https://github.com/rohitcode005/CCTV-Fight-Prediction-end-to-end-pipeline-dvc-aws-deploy.git

2. Create Virtual Environment

python -m venv venv

3. Install Dependencies

pip install -r requirements.txt

4. Run Setup Script

python setup.py install

5. Place Kaggle API Key

mkdir ~/.kaggle/
mv kaggle.json ~/.kaggle/

Setup and Running the Pipeline

1. Initialize DVC

dvc init

2. Run the Pipeline

dvc repro

or

python run_pipeline.py

Configuration Options

• MLFLOW_TRACKING_URI: https://dagshub.com/rohitcode005/CCTV-Fight-Prediction-end-to-end-pipeline-dvc-aws-deploy.mlflow
• MLFLOW_TRACKING_USERNAME: rohitcode005
• MLFLOW_TRACKING_PASSWORD: (Leave blank)

Additional Notes

• This project uses Keras and MLflow for tracking model accuracy and hyperparameter tuning.
• This project uses DVC to automate the workflow.
• This project is designed to be deployed on AWS using DVC.