📸 Image Authentication using Histogram Shifting and Arnold’s Cat Map

📌 Project Overview

This project presents an approach for Image Authentication based on the Histogram Shifting Method using Arnold’s Cat Map. The primary objective is to embed a watermark in images securely, ensuring their authenticity and integrity, which is especially useful for medical images. 🏥🔒

👨‍💻 Authors

• Niladri Ghosh

🏫 Institution

Ramakrishna Mission Residential College (Autonomous), Narendrapur, Kolkata - 700103

🚀 Features

• ✅ Watermark Embedding & Extraction: Secure image authentication through reversible data hiding.
• 🎨 Histogram Shifting Technique: Ensures imperceptible watermarking without significant image degradation.
• 🌀 Arnold’s Cat Map: Used for scrambling the watermark for enhanced security.
• 📊 Performance Metrics: Evaluates Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Normalized Cross Correlation (NCC).
• 🛡️ Attack Resistance: Resilient to image processing attacks like histogram equalization, median filtering, Gaussian noise, salt & pepper noise, cropping, JPEG compression, etc.

🛠️ Methodology

1️⃣ Embedding Process

1. 🖼️ Convert the image using Haar Wavelet Transform.
2. 🔍 Divide the image into LL, HL, LH, and HH sub-bands.
3. 🔄 Apply Arnold's Cat Map to scramble the watermark image.
4. 🔑 Encode the watermark using SHA-256 hashing.
5. 📥 Embed the scrambled watermark into the HH sub-band and the SHA-256 hash into LL sub-band using Histogram Shifting.
6. 🔄 Perform Inverse Haar Transform to reconstruct the stego image.

2️⃣ Extraction Process

1. 🔄 Apply Haar Transform on the stego image to extract LL and HH components.
2. 🖼️ Retrieve the watermark from HH sub-band and extract the SHA-256 hash from LL sub-band.
3. 🔄 Reverse Arnold’s Cat Map to recover the original watermark.
4. ✅ Verify watermark integrity using SHA-256 validation.

📊 Experimental Results

• 📂 Dataset Used: USC SIPI Image Dataset
• 📈 Average PSNR: 60.866 dB
• 🔢 NCC after Salt & Pepper Noise (0.001 density): 0.89
• 🛡️ Performance: Image authentication works well under normal conditions but is affected by aggressive attacks.

📦 Dependencies

Ensure the following Python libraries are installed:

pip install numpy opencv-python scipy

▶️ Usage

1. Run the attack script: Open the desired attack notebook (e.g., Attack/Median Blur_3/Average Filtering.ipynb) and run all cells to apply the attack on the images.

2. Run the extraction script: Open the Without Attack/code no attacks.ipynb notebook and run all cells to extract the watermark from the attacked images.

🏁 Conclusion

This project successfully implements a robust image authentication technique using Histogram Shifting and Arnold’s Cat Map. The method is effective for secure watermark embedding, but its resilience against aggressive attacks can be improved.

---