Passive Copy-Move Forgery Detection Using Halftoning-based Block Truncation Coding Feature

International Conference on Mathematics: Education, Theory and Application

数学;教育

Harjito, Bambang^1 ; Prasetyo, Heri^1

Informatics Department of Mathematics, Natural Sciences Faculty, Sebelas Maret University, Indonesia^1

关键词: Block truncation coding;    Compressed datum;    Copy-move forgeries;    Copy-move forgery detections;    Frequency threshold;    Image features;    Overlapping images;    Tampered images;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/855/1/012016/pdf
DOI  :  10.1088/1742-6596/855/1/012016

学科分类：发展心理学和教育心理学

来源: IOP

【 摘 要 】

This paper presents a new method on passive copy-move forgery detection by exploiting the effectiveness and usability of Halftoning-based Block Truncation Coding (HBTC) image feature. Copy-move forgery detection precisely locates the large size or flat tampered regions of an image. On our method, the tampered input image is firstly divided into several overlapping image blocks to construct the image feature descriptors. Each image block is further divided into several non-overlapping image blocks for processing HBTC. Two image feature descriptors, namely Color Feature (CF) and Bit Pattern Feature (BF) are computed from the HBTC compressed data-stream of each image block. Lexicography sorting rearranges the image feature descriptors in ascending manner for whole image. The similarity between some tampered image regions is measured based on their CF and BF under specific shift frequency threshold. As documented in the experimental results, the proposed method yields a promising result for detecting the tampered or copy-move forgery regions. It has proved that the HBTC is not only suitable for image compression, but it can also be used in the copy-move forgery detection.