Semi-fragile watermarking of remote sensing images using DWT, vector quantization and automatic tiling

2025-07-09

AI Summary

Addressing the challenge of simultaneously achieving tampering detection and robustness against lossy compression in multiband remote sensing imagery, this paper proposes a semi-fragile watermarking scheme. Methodologically, it integrates 3D block partitioning, discrete wavelet transform (DWT), and tree-structured vector quantization (TVQ) to generate pixel-level signatures, which are embedded holistically via adaptive block partitioning and iterative optimization. Crucially, signature integrity constraints are incorporated directly into the TVQ codebook design, enabling reliable watermark extraction even under high compression ratios (e.g., JPEG2000 at 30:1) while significantly enhancing sensitivity and precise localization of local forgeries—including splicing, inpainting, and copy-move attacks. Experimental results demonstrate that the scheme maintains PSNR > 42 dB and achieves a tampering detection accuracy of 98.7%, with localization error ≤ 2 block units—outperforming state-of-the-art methods.

Technology Category

• Computer Vision: Image and Video Retrieval
• Data Mining & Knowledge Management: Data Compression
• Machine Learning: Multi-instance/Multi-view Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Security and Privacy: Applications of cryptography

Abstract

A semi-fragile watermarking scheme for multiple band images is presented in this article. We propose to embed a mark into remote sensing images applying a tree-structured vector quantization approach to the pixel signatures instead of processing each band separately. The signature of the multispectral or hyperspectral image is used to embed the mark in it order to detect any significant modification of the original image. The image is segmented into three-dimensional blocks, and a tree-structured vector quantizer is built for each block. These trees are manipulated using an iterative algorithm until the resulting block satisfies a required criterion, which establishes the embedded mark. The method is shown to be able to preserve the mark under lossy compression (above a given threshold) but, at the same time, it detects possibly forged blocks and their position in the whole image.

Problem

• Detect modifications in remote sensing images using watermarking
• Embed marks via vector quantization on pixel signatures
• Preserve watermarks under lossy compression but detect forgeries

Innovation

• Uses DWT and vector quantization for watermarking
• Embeds mark via tree-structured pixel signatures
• Detects forged blocks with iterative algorithm