Blur Detection in Image Forensics Using Linear Correlation of Pixels

Based on linear correlation detection, in this paper a novel digital image forensic scheme is proposed, wherein whether an image is tampered or not may be determined using blue operation. In order to detect blur, which can reveal the trace of tampering, we analyze the linear correlation of image pixels with shortest distance clustering approach. Experimental results show that the proposed method could reveal blurred regions that indicate possible tampering without any embedded information such as watermarking technique.