International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017
p-ISSN: 2395-0072
www.irjet.net
An Efficient Image Forensic Mechanism using Super pixel by SIFT and LFP Algorithm SHAHABUDDIN.S.K 1, DR. A.R.ASWATHA2 Mtech Student, 4th semester, Department of Telecommunication, Dayananda Sagar College of Engineering, Bangalore -78, Karnataka,India 2Professor and Head, Department of Telecommunication, Dayananda Sagar College of Engineering, Bangalore-78, Karnataka,India ---------------------------------------------------------------------***-------------------------------------------------------------------1
Abstract—Image forgery is a very significant and challenging area in the field of image processing and its applications. One of
the very commonly used image forgery technique is the copy-move technique which is not only a common technique but also a challenging technique. Initially a wavelet decomposition method is applied in order to compute the size of super pixels which is further applied to segmentation block. Post segmentation the feature extraction and feature matching is performed using suitable methods. Consequently region growing is performed in order to detect the forged area in the image. Finally the region of forgery is detected and evaluated using the measures of precision, sensitivity, specificity and F1 score. Experimental results show an increased performance in the above defined measures. Key-Words: Image Forgery, Copy-Move Method, Wavelet Decomposition, Segmentation, Region Growing
1. Introduction In the present day of image processing applications, the sophistication and complexity of images has increased to a considerable extent, which provides a broader scope in variety of areas ranging from military applications to biomedical applications. However, the increase in altering (editing) the image has linearly increased, with the commercialization of image editing tools, the issue has only deepened to the extent where the credibility and authenticity of the image is questionable. One practical application highlighting this issue involves the editing of photography in a crime scene; the aggregator could alter the content in the image deliberately in order to hide the key evidence which might otherwise be visible in the respective image. Image forgery is defined as a manipulation of digital image in order to conceal or alter the meaningful or useful information present in the image. The first person to have successfully altered the image was in 1840 by a man named Hippolyta Bayard who edited a picture of himself. However, the extent of digital forgery in the image does not differ much from its original image which makes it difficult for the interpreter to identify the element of forgery in the edited image. The Copy-Move technique is considered as one of the most commonly used image tampering technique. It is also considered as one of the most difficult technique. This is due to the factor that the manipulation of image content is within the image itself. In this technique, a part of the image is deliberately covered by copying a part of the image and pasting (Moving/ superimposing) it in another place in the same image. The intention of implementing this technique is to alter the image by hiding a portion/ part of an image. Tu and Dong [11] have significantly explained the image feature matching and also the histogram equalization mechanism. The algorithm implemented for features matching of an image will provide the better efficiency with the image scale, rotation, illumination difference for practical applications. The image acquisition process includes various conditions like posture, position and camera performance etc., for the image area. The projective distortion of the image area will affect the accuracy in FP matching/extraction. The author has given the preprocessing mechanism to enhance the significant features of the image by using the histogram equalization. Later the image is subjected to ASIFT and SIFTS algorithms for matching and extraction. The experiment conducted by the author gives the improved way to get more number of feature points by histogram equalization. The color spaces evaluation to detect the feature point of image for the matching application is explained in Sirisha and Sandhya [12]. The study of the authors includes the inspection of the color information according to the FP detection. These color information’s are expressed as HSV, XYZ, RGB, LAB, YIQ, CMY, CbCr and opponent. Later more relevant color space will be problem and hence the processing of color image is done. The image matching is done with the Hybrid color space while the Principal Component Analysis (PCA) is used for feature selection. The experiment is conducted by keeping the total repeatability measurement and total feature points for evaluation.
© 2017, IRJET
|
Impact Factor value: 5.181
|
ISO 9001:2008 Certified Journal
| Page 2806