BIIGA: Bioinformatics inspired image grouping approach

Abstract

Initial work on image phylogeny used different approaches like the minimum spanning tree etc. The less investigated attempt is a bioinformatics-inspired approach for image phylogeny. The aim of this paper is to bridge this gap by generating image phylogeny with the concept of phylogenetic trees in bioinformatics. These trees were developed by using multiple aligned DNA sequences of original or multiple print-scan (MPS) degraded variants of watermarked (W) and non-watermarked (NW) images. Experimental results disclosed the viability of a proposed novel approach and effectively grouped original or MPS degraded variants of the W or NW images. The proposed method’s claims may revolutionise our knowledge of degraded (D) or non-degraded (ND) W and NW image grouping. It may direct to a new era of phylogenetic tree-based W or NW degraded or original images for developing the next generation of degraded (by MPS) W and NW grouping softwares. To the best of our knowledge, this is the first time such a method has been employed for image grouping. Our contributions are: (a) biologically-based image encoding to DNA letters (b) multiple sequence alignment (MSA) of DNA-encoded images (c) phylogenetic tree generation to group MPS-degraded W or NW images.