Title :
A method for identifying computer images and real images
Author :
Guo, Ke ; Wang, Rangding
Author_Institution :
CKC Software Lab., Ningbo Univ., Ningbo, China
Abstract :
Based on the differences of pattern noise between real images and computer images, this paper advanced a new method which combined the SNR features with the higher order characteristics of the predicting error images. Where the SNR features consist of MSE, SNR and PSNR features between original images and modified images which were got by add-noising and de-noising for the original images. Experimental results show that this algorithm´s recognition rate can get 94.33% in the Columbia Image Dataset [1].
Keywords :
feature extraction; image recognition; Columbia image dataset; SNR features; computer image identification; pattern noise; real image identification; recognition rate; Computational modeling; Computers; Feature extraction; Image color analysis; Prediction algorithms; Signal to noise ratio; Add-noising and de-noising; SNR features; wavelet transform;
Conference_Titel :
Electronics, Communications and Control (ICECC), 2011 International Conference on
Conference_Location :
Ningbo
Print_ISBN :
978-1-4577-0320-1
DOI :
10.1109/ICECC.2011.6066370
