EchoMark is a forensic-grade watermarking system designed to ensure the authenticity and integrity of digital surveillance media — images, audio, and video. Built for law enforcement and digital forensics, it provides imperceptible, tamper-resistant, and reversible watermarking modules with performance-tested results.
You can download the latest version of the EchoMark installer via:
- 🔒 Image Watermarking (Reversible)
Spatial domain embedding using a weighted-sum method for lossless original recovery. - 🔊 Audio Watermarking (Robust)
FFT-based frequency embedding resistant to pitch, volume, and compression. - 🎥 Video Watermarking (Tamper Detection)
Frame-wise watermarking with SSIM comparison for forgery localization. - 📊 Performance Benchmarks
- PSNR > 40 dB (imperceptible)
- Extraction Accuracy: ~95%
- BER ≈ 0.01
- Tested across 65 multimedia samples
- 🔐 Metadata & Signature Encryption
- 📈 Visualization
- Flowchart of system modules
- Comparison charts for tampered vs. untampered samples
- 💻 GUI Application
Built with MATLAB front-end and Python watermarking back-end.
| Frontend | Backend (Toolside) |
|---|---|
| React.js | Python (for modules) |
| Bootstrap 5 | OpenCV, NumPy |
| React Router | FFT / SSIM Algorithms |
| Chart.js | MATLAB (GUI Launcher) |
| AOS (Scroll Reveal) |
| Media Type | PSNR (±SD) | BER (±SD) | Accuracy (±SD) | Samples |
|---|---|---|---|---|
| Image | 42.87 ± 1.94 dB | 0.012 ± 0.004 | 97.4% ± 1.3% | 30 |
| Audio | 38.10 ± 2.32 dB | 0.019 ± 0.005 | 94.6% ± 2.1% | 30 |
| Video | 40.26 ± 2.05 dB | 0.015 ± 0.003 | 95.8% ± 1.7% | 30 |
EchoMark contributes toward:
- SDG 16: Peace, Justice and Strong Institutions
Enables secure, traceable, and verifiable digital surveillance evidence. - SDG 9: Industry, Innovation and Infrastructure
Supports innovation in digital forensic infrastructure.
The watermarking modules were evaluated on real-world attacks including:
- JPEG compression (up to 60%)
- Audio bitrate downsampling (320 kbps → 128 kbps)
- Video cropping and Gaussian blur
- Frame tampering, noise, and rotation
🧬 Result: EchoMark's embedded watermarks consistently survived with high accuracy.
Watermark survived all cases with measurable resilience.
| Name | Role |
|---|---|
| Divyanshu Ranjan Singh | System Architect, Developer |
| Aastha | Research & Implementation |
| Merdul Sharma | UI & Video Module |
| Aman | Audio Module & Testing |