This project aims to design and implement an AI-based intelligent camera decision-making system to address challenges in traditional surveillance methods. The goal is to process video data, identify crucial events, and make real-time informed decisions. The system addresses issues such as human fatigue, delayed responses, and the need for continuous monitoring.
- A cost-effective approach leveraging existing infrastructure.
- Upgrade camera firmware for compatibility.
- Implement adaptive machine learning for continuous improvement.
- Utilizing the Mediapipe framework for face and body pose recognition, hand tracking, etc.
- Implementing deep learning models (CNN) for object (arms) detection.
- Object detection using OpenCV.
- Mobile app integration for instant updates on suspicious activity.
- Location, photo, and timestamp of the incident provided with the alert.
- Email alerts for maintaining a record of the activity.
- Photo proof sent along with the alert to avoid misjudgments.
- High-resolution cameras for advanced video capture.
- GPU for accelerated deep learning algorithms.
- Multicore CPU, sufficient RAM, and high-speed storage.
- Robust network infrastructure for seamless data transfer.
- Programming Languages: Python, C++, Kotlin
- Frameworks: TensorFlow, PyTorch, CNN, LSTM, SVM, etc.
- Tools: OpenCV for computer vision, Mediapipe for body postures and hand tracking
- Cloud Services: Google Cloud Platform
- Kshitiz Agrawal: Integration of mobile and sensors for data collection.
- Aarohi Saxena: Machine Learning & Research.
- Rachit Goyal: Sensors integration and IoT.
The proposed AI-based intelligent camera decision-making system overcomes limitations in traditional surveillance methods, offering real-time analysis, quick decision-making, and continuous monitoring. The interdisciplinary approach makes it suitable for integration in police stations, empowering authorities with quick and authentic crime case information.
This README provides an overview of the AI-based system, its components, and the team's responsibilities. For a detailed guide on installation and usage, please refer to the project documentation.