Forgery detection in digital images using Self Organizing Map

Abstract

In this paper, a passive scheme to determine the tampered region in digital images is proposed. The proposed scheme consists of splitting the image, estimation of quantization table from the peaks of the histogram of DCT coefficients, clustering the blocks using Self Organizing Map (SOM) which detects the tampered region. The variation resulting from the inconsistency of quantization table is utilized to detect tampered regions efficiently. The results show that the proposed scheme, effectively detects the tampered region.