A duplicate image detection scheme using hash functions for database retrieval

This paper presents a new duplicate image detection scheme that employs hash functions to cope with the problem resulting from the large image database containing the fingerprints extracted from the images. Moreover, it utilizes the extracted features of the host image to generate a fingerprint, which is then classified in the database systematically. When the authentication of suspect images is needed, the scheme only compares the matches from the same image to decrease the comparison time. The scheme uses the feature vector from the fingerprints to build a hash pool structures, and utilizes the phase angle calculated from the feature vector to query the probably duplicate images in the image database. The experimental results show that the presented scheme performs well according to the recall and precision rates, hence demonstrating the effectiveness of the presented scheme.