Detecting Forged Scan of Educational Certificates Using a New Feature Set Matching Algorithm

Abstract

Detection of forge scan certificates which are used during college admissions are done using scan copies from other genuine resources and materials and resources applying Photoshop and other image processing tools. This kind of certificates from different resources are used for creating scan certificate copies using Photoshop and giving that information to the name of other candidates. This situation leads to a point whereupon digital forgery can compromise the authenticity of the original documents. Using fake certificates to get jobs is worst as the eligibility of the candidate cannot be verified by administrative authorities in educational institutes. Nowadays many research oriented tools are used for detecting copied sections from different documents. In order to enhance trustworthiness and authenticate images, we have defined a feature point matching algorithm to identify forgery done in inaccurate manner. In our proposed study we have applied by an image processing algorithm to combine feature point matching tools and adaptive segmentation to identify suspected irregular patterns detected by the adaptive non-overlapping and irregular blocks and this process is carried out using the adaptive over-segmentation algorithm. The extraction of the feature points is performed by identification of the matching between each block and its features. The feature points are gradually replaced by using the super pixels in the proposed Forgery Region Extraction algorithm and then combine the adjacent blocks having similar color features to find out the merged regions for identification of the detected forgery regions.