|Title:||Image forensics by detecting blurred regions|
|Subject:||Hong Kong Polytechnic University -- Dissertations|
Image processing -- Digital techniques
Pattern recognition systems -- Mathematical models.
|Department:||Department of Computing|
|Pages:||ix, 85 leaves : ill. ; 30 cm.|
|Abstract:||To review and assay the genuineness of device-independent images, blur estimation by relative frequency and semantic features (BERFS)  is proposed. BERFS system is a tool that can detect the correctness of forged images. It introduces the concept of relative frequency feature which can show the difference between the blurred region and non-blurred region. It is claimed that estimation of blurred regions can be made with high precisions . In this project, a detail investigation is carried out to understand how BERFS system can work well on the forged images. Parameter analysis is a main focus of this study. In the BERFS system, there exist several input parameters and our objective is to analyze their sensitivity to the overall performance. Besides, the semantic features contained in the image are investigated. Although there are some suggested methods to select the regions that contain similar semantic regions  , they are still estimation based and it is possible that some selected regions contain different semantic regions. If the size of blur region increases, the performance of BERFS may be changed also. In addition, we evaluate the image content that will affect the performance of image forensics based on BERFS. Experiments were conducted to evaluate the BERFS system with respect to the concerns stated above and the results shield light on fine tuning the BERFS method.|
|Rights:||All rights reserved|
Files in This Item:
|b23526415.pdf||For All Users (off-campus access for PolyU Staff & Students only)||7.66 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: