A Noval Approach for the Detection of Forgery to Improve Classification Accuracy

Authors

Jagminder Kaur, Raman Kumar
Department of Computer Science and Engineering, I K Gujral Punjab Technical University, Kapurthala, Punjab, India.

Abstract

Objectives: To improve the peak signal to noise ratio to detect the forgery within video frames using color scaling mechanism. To improve classification accuracy of forgery detection using tangent based approach. To compare tangent based approach with bounding box and proposed mechanism. Methods: Tangent based bounding box mechanism is applied in order to extract the features from the video frame to detect forgery. Result is expressed in the form of peak signal to noise ratio and classification accuracy. The peak signal to noise ratio is improved using mean shift contrast enhancement strategy and classification accuracy using tangent based approach. Findings: Overall, the result shows an improvement by 9% accuracy Novelty: The static video frames extraction can be accomplished easily using only bounding box mechanism but in case video frame is dynamic then extraction may not be possible using plain bounding box mechanism. To overcome the issue tangent based approach is hybridised with the bounding box mechanism. This process detects the forgery at higher classification accuracy as compared to existing approaches.