Comparison of near-duplicate image matching
Lihui Chen, Fred W. M. Stentiford · 2006
29-30 Novemeber 2006. Near-duplicate image detection requires the matching of slightly altered images to the original and will assist in the detection of forged images. Much effort has been devoted to visual applications that require effective image signature and similarity metrics. This paper presents an attention based similarity measure in which only very weak assumptions are imposed on the nature of the features employed. This approach generates the similarity measure on a trial and error basis and has the significant advantage that matching is based on an unrestricted competition mechanism that is not dependent upon a priori assumption regarding the data. Efforts are expended searching for the best feature for specific region comparisons rather than expecting that a fixed feature set will perform optimally over unknown patterns. In this paper colour and texture-based signatures are extracted to compare the presented method in the context of near-duplicate image matching, and results are reported on the BBC open news archive.