Compressive sensing based watermarking as a security layer for computational imaging applications

The evolution of digital imaging since its early days has mostly been driven by hardware advancements. Computational imaging is a recent alternative to traditional imaging, that focuses on the use of digital processing algorithms. One notable example is the single pixel imaging and its recent developments in the area of lensless imaging. It uses image acquisition based on the compressive sensing (CS) theory, resulting in measurement data that use less storage space. However, it requires more computational resources in order to reconstruct the image and make it available for the end user. The CS theory also opens new research opportunities regarding the security of digital image processing, storage and distribution. This paper presents a privacy enabling architecture for a real-life scenario of image acquisition using the single pixel camera. A reversible watermark is inserted to the measurement data before storage and distribution.