Title :
A novel steganalysis of LSB matching based on kernel FDA in grayscale images
Author :
Hu, Lingna ; Jiang, Lingge ; He, Chen
Author_Institution :
Dept. of E.E., Shanghai Jiao Tong Univ., Shanghai
Abstract :
To detect presence of LSB matching blindly, a novel steganalysis is proposed. First, image segmentation is employed to separate image into different domains. Second, statistic property of node degree for minimum spanning tree (MST) in random domain is analyzed. And third, local image complexity is proposed to describe concrete domain situation, and image features are also extracted accordingly. Simulation results demonstrate that the proposed algorithm can achieve higher detection probability than existent ones on both uncompressed and compressed image formats, especially low embedding rate.
Keywords :
computational complexity; cryptography; data encapsulation; feature extraction; image matching; image segmentation; statistical analysis; trees (mathematics); LSB matching; compressed image formats; detection probability; grayscale images; image features extraction; image segmentation; kernel FDA; local image complexity; minimum spanning tree; random domain; statistic property; steganalysis; Color; Gray-scale; Histograms; Image segmentation; Intelligent networks; Kernel; Neural networks; Signal processing; Steganography; Training data; Kernel FDA; LSB steganalysis; blind detection; image segmentation; local image complexity;
Conference_Titel :
Neural Networks and Signal Processing, 2008 International Conference on
Conference_Location :
Nanjing
Print_ISBN :
978-1-4244-2310-1
Electronic_ISBN :
978-1-4244-2311-8
DOI :
10.1109/ICNNSP.2008.4590412
