Authentication of compressive sensing based image content

Más información sobre el libro

Compressive Sensing (CS) has transformed sampling methods by selecting suitable sensing domains, constraining source information characteristics, and utilizing nonlinear solvers for recovery. CS-based Imaging (CSI) cameras can generate an N pixel image using only M measurements, where M is significantly less than N, and they are not limited by sensor bandwidth when the target image is k-sparse or compressible. To ensure the integrity and authenticity of CS-based images, an authentication mechanism is essential. This dissertation introduces an authenticated CSI system utilizing the Compressive Sensing based Message Authentication Code (CSMAC) mechanism, which is integrated into the imaging process through a redundant secure matrix. The extraction mechanism reduces data size and supports a restricted tolerance property. A pre-trained threshold allows the verifier to accommodate a specific level of recovery noise while detecting content modifications. Additionally, an authenticated encrypted CSI mechanism is proposed to combine authentication, encryption, compression, and sensing in a single step. Given the subpar data compression performance of CS-based images, this study suggests improvements for the authenticated encrypted CSI mechanism. Simulated results demonstrate that CSI is computationally secure for confidentiality, sensitive to content-based tampering, and achieves approximately a 20% lower bit-rate for acceptable image