Robust image hashing with dominant DCT coefficients

Image hashing is an emerging technology of multimedia processing. This paper proposes a robust image hashing with dominant discrete cosine transform (DCT) coefficients. The proposed hashing converts the input image to a normalized image, divides it into non-overlapping blocks, extracts dominant DCT coefficients in the first row/column of each block to construct feature matrices, and finally conducts matrix compression by calculating and quantifying column distances. Many experiments are conducted and the results show that the proposed hashing is robust against normal digital operations and has desirable discrimination. Receiver operating characteristics (ROC) curve comparisons indicate that the proposed hashing is better than some notable image hashing.