Open AccessImage Forgery Detection Using Dct And Quantization Matrix Techniques
Author(s) -
Siby Abraham,
Anisha P Rodrigues,
Roshan Fernandes
Publication year - 2019
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.f8883.088619
Subject(s) - discrete cosine transform , quantization (signal processing) , jpeg , artificial intelligence , computer vision , computer science , image (mathematics) , block (permutation group theory) , image processing , image quality , transformation matrix , transformation (genetics) , pattern recognition (psychology) , mathematics , biochemistry , chemistry , physics , geometry , kinematics , classical mechanics , gene
In today’s era the image has become useful for communication purpose. But due to the development of software and various techniques it is possible to change images in adding or removing essential feature from it without leaving a clue of real image. It is not easy for the common people to identify whether the image original or tampered. In order to avoid this problem, forgery detection came into existence. Detection of forgery refers to task of image processing to identify that the images are unique or tampered. Several techniques have been used in order to detect the forgeries from the forged image, but this issue has not yet solved. In order to solve these issues we have used Discrete Cosine Transformation (DCT) and quantization matrix techniques for identifying forged areas of image, where the quality of image is not reduced. The Discrete Cosine Transformation (DCT) is used in order for characterizing the overlapping blocks and quantization matrix is used to compress DCT values and gives both highly compressed and best decompressed image quality. Here we use block matching algorithm. This algorithm one of the most frequently used for detecting image which is duplicate. This proposed work also supports for different kinds of images such as JPEG, JPG or PNG of any size it can be either mxn or nxn.