On-line Hybrid Perceptual and Cryptographic Image Hashing System

Vision systems became safety critical in modern 'Advanced Driver Assistance Systems', where software-controlled high-definition cameras and digital signal processors are the core components. Software verification using ‘Hardware-In-Loop’ is a growing need as a result. In this paper a new hybrid perceptual hashing system for on-line digital video images processing is presented. The system can detect frozen images, mono-color images and may be used in image comparison as well. The design uses minimal hardware and storage while maintaining a high operational frequency. The system performance is analyzed and functionally verified on an HDL implementation using an FPGA.