Author:
Dmitriev E A,Myasnikov V V
Abstract
Abstract
This paper presents an approach for feature extraction using the equivalence relation. The equivalence relation is formed using features permutations of primary data. Every feature of new representation matches to optimal permutation, that minimizes some loss function. A loss function choice depends on the kind of problem. In this paper, we consider copy-move forgery detection problem and demonstrate the effectiveness of our approach in solving this problem.
Subject
General Physics and Astronomy
Reference10 articles.
1. Method for detection of vehicles in digital aerial and space remote sensed images;Myasnikov;Computer Optics,2012
2. Geometry of sign representation of images and its application to noise resistance investigation;Karkishchenko;Intelligent Data Processing: Theory and Applications,2012
3. Constructing efficient linear local features in image processing and analysis problems;Myasnikov;Automation and Remote Control,2010
4. Performance evaluation of texture measures with classification based on Kullback discrimination of distributions;Ojala;Proceedings of 12th International Conference on Pattern Recognition,1994