PLAGIARISM DETECTION IN ACADEMIC DOCUMENTS USING COMBINED DCT+WAVELET+INVARIANT TRANSFORM BASED COPY PASTE DETECTION

The paper describes in brief about the plagiarism detection and various works done by different authors and the final section describes about the proposed method’s efficiency as it eradicates and solves all the above mentioned occurring problems in the bas e technique i.e. Linear Programming technique. The importance of the technique is that it makes use of a method which detects and extracts more features as compared to method used in base technique. In comparison to the base technique the proposed will det ect plagiarism in documents for text copied from other documents, for edited photo and images from various documents. It will identify accuracy of plagiarism detection in document images and will further compare the proposed work with the previous techniqu es under defined parameters. In the proposed method, the images document will be selected to process to cut and paste. Further the improvement filters will be applied if needed. Next step involves application of hybrid similarity and arte - fact detection us ing multi - temporal filtering using DCT analysis and classification using SVM and feature extraction that results in segmenting the image into parts. The proposed method uses the extracted feature for matching with documents suspected to be copied from and after matching test the accuracy, efficiency, time of the processing algorithm we further use image documents with / without distortions for testing fluency of the algorithm.