🕵️‍♂️ StegoTester v3.1

The Ultimate Steganography Analysis & Detection Tool Powered by Hybrid AI Architecture (Gatekeeper + Detective)

StegoTester is a professional-grade forensic tool designed to detect hidden data within digital media and benchmark steganography algorithms.

Unlike traditional tools that rely solely on simple math, StegoTester v3.1 employs a Hybrid Detection Engine that combines rigorous statistical analysis with trained Machine Learning models.

🧠 AI Training Module: The source code for model training, dataset generation, and feature extraction used in this project can be found in the StegoTester-AI-Training repository.

---

🚀 New in v3.1: The Hybrid Engine

This version introduces a revolutionary detection architecture:

1. 🛡️ The Gatekeeper (Audio)

A zero-latency mathematical guard for audio files.

• Calculates the LSB Transition Rate instantly.
• If the rate exceeds the 0.455 threshold, it flags the file as CRITICAL STEGO immediately, bypassing the AI for maximum speed.
• Result: 99.9% accuracy on standard LSB attacks.

2. 🕵️‍♂️ The Detective (AI Models)

Files that pass the Gatekeeper (complex audio) and all images are sent to our trained Machine Learning models (Random Forest / Gradient Boosting).

• Analyzed Features: Entropy, Edge Density, MFCC, Spectral Flatness, LSB Correlation.
• Result: Detects subtle anomalies (like Echo Hiding or Texture Masking) that human senses miss.

---

📸 Screenshots

Audio Analysis & AI Detection

The Gatekeeper in action: Detecting LSB noise with 1.0 confidence, while analyzing complex Stego attacks.

Image Analysis Metrics

Comprehensive breakdown of MSE, PSNR, and SSIM metrics alongside AI probability scores.

---

🛠 Features

📊 Comprehensive Metrics

Benchmarking standards for academic and forensic use:

• Image: MSE, PSNR, SSIM, LPIPS (Perceptual).
• Audio: SNR, LSD (Log-Spectral Distance), MAE.
• Payload: BER (Bit Error Rate) verification for extracted data.

🧠 Smart Matching

Automatically pairs Original and Candidate files using pHash (Images) and Spectral Fingerprinting (Audio). No need to rename files manually!

🖥️ Modern Reporting

• Visual Comparison: Interactive "Before/After" slider.
• PDF Exports: Professional reports with Executive Summaries, Risk Pie Charts, and detailed findings.

---

📉 Interpreting AI Scores

The AI_DETECTION metric represents the probability (0.0 - 1.0) of a file containing hidden data.

Score Range	Verdict	Explanation
0.00 - 0.50	🟢 SAFE	File appears clean.
0.50 - 0.80	🟠 SUSPICIOUS	High complexity or noise detected. Could be a subtle stego attack OR a naturally complex file (e.g., highly textured image).
0.80 - 1.00	🔴 CRITICAL	Strong statistical evidence of tampering. Almost certainly contains hidden data.

---

📦 Installation & Setup

1. Clone the Repository:

   git clone https://github.com/umitkrkmz/stegotester.git
   cd stegotester

2. Install Dependencies:

   pip install -r requirements.txt

   (Note: Libraries like Librosa and Torch may take a moment to install)

3. Model Check: Ensure trained models are present in the models/ directory:

   • models/stego_model_image.pkl
   • models/stego_model_audio.pkl

   Note: Visit the StegoTester-AI-Training repo to see how these models were trained.

4. Run the Application:

   python main.py

---

📜 License

Distributed under the MIT License. See LICENSE for more information.