A Tool for Video Forgery Detection in Video Sequences

Saminda Premaratne, A.S.C.I Piyumalie, L.L.T Pubudini, B.U.S Senarathne, K.M Thilakarathne


Because of the vast development of the video editing tools, the digital videos are playing main role in todays’ context. Because of the variety of techniques available, it is very easy to alter and edit original content of a video. So paper describes about implemented system for video surveillance which could apply for tamper detection. As an inputs system is getting the videos which are needed to be tested. System could use, for law enforcements, CCTV (Closed-circuit television) videos, for news broadcasting. Mainly this system is considering developing a system for analysing shadow variation with the noise and system recognize an object and extract shadow from the object and for detect moving objects and track objects. After this approach system will identify pattern variations in shadow, motions and create a model by using SimpleKMeans approach as a clustering technique in data mining and frame removed edited videos by using Naïve Bayes classification algorithm in data mining. According to the pre identified model, user input videos will be identified as tampered or not. The model creation was done for 3 CCTV s. From each CCTV system has taken 20 videos .And implemented the model for 5 noise levels, kernel size11, 15,25,55,75 for shadow detection and frames were removed in-between original videos for different time levels. According to model, input video from predefined CCTV gives noise level of the video and error rate of each noise level .And objects are tracked and according to motion vector, the variance of the objects will be calculated. This result is used in data mining and according to that prediction happens.

Keywords: SimpleKMeans, Naïve Bayes, Data Mining, Motion Vector, CCTV, Weka

Full Text:


DOI: https://doi.org/10.26483/ijarcs.v6i2.2425


  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science