Develop a prototype for a liveness detection system that uses videos from camera selfies. Your solution should be secure and robust against basic spoofing attacks.
- Liveness Detection Algorithm: Create a machine learning model to distinguish between live individuals and non-live entities (e.g., photos, video replays, masks). Describe your model selection, feature engineering, and training process.
- Adversarial Machine Learning: Implement and test your model against adversarial attacks designed to bypass liveness detection. Detail your approach for generating adversarial examples and how you enhance your model's robustness against these attacks.
- Security Measures: Address the security aspects to protect the integrity of the liveness detection process. Discuss potential system vulnerabilities and your strategies to mitigate them.
- Testing and Validation: Describe the methodologies you would use to test the liveness detection and its resilience to adversarial attacks. Include performance metrics to evaluate your system’s effectiveness.
- Code Repository: Provide a link to a Git repository containing all source code, with instructions for setup and execution, both, the liveness model and the adversarial attack.
- Include a comprehensive report detailing your methodologies, choice of technologies, the rationale behind key decisions, system architecture and data flow diagrams, and a summary of challenges encountered during development and their solutions.
- Technical Proficiency: Efficiency and quality of the code, appropriate use of algorithms and technologies.
- Innovation and Creativity: Originality in approach and problem-solving capabilities.
- Depth of Understanding: Clarity in explaining and justifying design and implementation choices.
- Security Mindset: Awareness of potential security threats and the effectiveness of implemented countermeasures.