An Anti-JPEG Compression Image Perceptual Hashing Algorithm

Perceptual hashing schemes were widely applied to image content authentication fields in recent years. However, most hashing algorithms were difficult to differentiate malicious attacks from normal signal processing operations, such as image compression, enhancement. As well known, digital images are JPEG compressed mainly by quantifying the DCT coefficients. The higher frequency DCT coefficients are removed, while the lower values are protected in this process. This paper will propose a novel image perceptual hashing method based on image DCT low frequency coefficients which not only can achieve authentication for the JPEG image, but also can detect some malicious manipulations effectively, including tampering, noise. Also, Theoretical and experimental results are used to demonstrate the effectiveness of this algorithm.