01Overview
Dignity Shield AI is a deepfake detection awareness system specifically designed to combat synthetic media harassment and protect youth digital identities.
02The Problem
Existing enterprise cyber forensics tools are prohibitively expensive and fail to provide actionable student-friendly privacy defense guidance.
03The Idea & Thesis
A lightweight web analyzer that inspects biological micro-signals and produces an explainable authenticity score.
04Process & Methodology
1
Requirements elicitation and domain constraint modeling
2
Architecture design and prototyping in TypeScript/Python
3
Stress testing, optimization and institutional deployment
05Technology Stack
Core Stack
PythonOpenCVPyTorchFastAPIReactTailwind CSS
06My Role & Contribution
- AI Researcher & Lead Developer
07Key Technical Challenges
Handling compressed social media video files where compression noise mimics generative artifacts
08Measurable Results & Impact
Milestone #1
Selected for National Finals at eRaksha Tech Avishkar
Milestone #2
Achieved 89.4% classification accuracy across synthetic face swap benchmarks
09Media Highlights
10Core Learnings & Reflections
Multimodal temporal signals outperform single-frame classifiers on compressed video
11External Links & Resources