Forgery detection using noise variance estimation and HOG features

In this book, a method is proposed to detect the forgery based upon noise estimation and hog feature extraction on NTSC color image i.e. YIQ color space. The YIQ colorspace is used because it splits the luminance and the color information. For forgery detection, the color information can be discarded as only the variance in the luma component is helpful. The image is first converted to YIQ color space, and then the block segmentation is performed on Y component of the YIQ image. Noise estimation and hog features are extracted from each block of the image. The method used for noise estimation is Principal component analysis (PCA) which estimates the noise as the smallest eigen value of the covariance matrix of the image block. An unsupervised clustering method is used to cluster the blocks of the image based upon noise and hog features combined together. Then, SVM classifier is used for refinement of the clustered blocks. The experimental results show that the proposed technique detects forged images more effectively as compared to previous method which is based only on noise variance estimation.

Autorentext

Savita Walia has received Masters of Engg. in I.T. in 2015 from Panjab University, Chandigarh, India and the bachelor's degree in computer science engineering, in 2012 from Punjab Technical University, Jalandhar, Punjab, India. Currently, she is pursuing the Ph.D. degree in the area of Digital Image processing from Panjab University, Chandigarh.