Image Perceptual Hashing for Content Authentication Based on Geometric Invariant Vector Distance

Image perceptual hashing is broadly applied in image content authentication, recognition, retrieval, and social media hotspot event detection. An authentication algorithm put forward based on the Itti visual saliency model geometric invariant vector distance. To begin with, preprocessed weighted by contourlet transform. After that, randomly divided into blocks, feature constructed calculating distance both Hu moment maximum singular value of random blocks. In end, quantized encrypted to generate ultimate hash. Experimental results illustrate that when threshold T = 70, true positive rate P T mathvariant="normal">P mathvariant="normal">R for duplicate images stands at 0.96574, while false id="M2"> mathvariant="normal">F different merely 0.0224, with total error reaching minimum (0.0566). Furthermore, AUC proposed 0.9951, which higher than comparison algorithms, indicating has better performance other state-of-the-art algorithms terms various content-preserving attacks.