A Biometric Features-Based Watermarking Encryption Approach for Secure Authentication of Visual Images

Abstract

Listen

Translate article

Due to the uniqueness of biometric features, it remains practical in generating cryptographic keys. In this study, a region-based encryption method is proposed, which can avoid the attacks on the whole encryption such as block attack and frame insertion attack. This method can protect important areas while encrypting images so that the image can maintain recognition and credibility even when attacked. It has certain universality and scalability, and can be applied to a variety of scenarios and systems. First, biometric features are introduced to extract keys from fingerprint key images through grayscale segmentation and gradient segmentation. Then, a watermarking addition algorithm is completed based on single encryption and reversible exchange encryption using hybrid vectors, and the image is encrypted using encryption algorithms. Next, the generated biometric key is embedded into the encrypted image, and a complete watermark encryption model named BKE is established by adjusting the singular value decomposition algorithm. Finally, invisibility, decoding accuracy, and stability analysis are completed based on publicly available data. The experimental results show that this method has high effectiveness and feasibility in image security authentication, and this method improves the security and authentication accuracy of images. This method has broad application prospects and can be used for image security authentication in various fields.