An image hashing for content authentication has been paid large attention from researchers. But, simultaneous achievement of robustness to geometric distortions, good discrimination, and identifying the areas of tampered regions in an image is still an open issue. To resolve the above issue, the proposed system includes ring partition with Gray level cooccurrence matrix ( GLCM), an exemplar-based saliency detection, and a blind geometric correction. First, the global features are extracted based on GLCM from rotations invariant regions, i.e., via ring partitions. Next, the local features are extracted using an exemplar-based saliency detection method. The two features are concatenated to form a final hash. At the time of image authentication, the geometric transformations are mitigated via a blind geometric transformation correction approach. The experiment results carried out on large standard image pairs show that the proposed provide better robustness, good discrimination, and identified the tampered areas. The efficacy of the proposed is shown using a true positive rate (TPR) and false positive rate (FPR).
Image Authentication System using Ring Partition and GLCM
2020-11-05
471718 byte
Conference paper
Electronic Resource
English
Use of operational software for GLCM missile procedures trainer
Tema Archive | 1983
|Using GLCM features in Haar wavelet transformed space for moving object classification
Wiley | 2019
|Using GLCM features in Haar wavelet transformed space for moving object classification
IET | 2019
|