FORENSIC VISION: DETECTING COPY-MOVE FORGERY IN DIGITAL IMAGES

Abstract

In the digital era, the authenticity of images has become increasingly vulnerable due to the ease of manipulation through sophisticated editing tools. Among various tampering methods, copymove forgery—where a region of an image is copied and pasted within the same image to conceal or duplicate objects—is one of the most prevalent and challenging to detect. Traditional manual inspection often fails due to the seamless blending techniques used, necessitating the development of advanced image forensic algorithms. This study introduces Forensic Vision, a framework for the detection of copy-move image forgery using a combination of feature extraction, block-based analysis, and similarity matching techniques. The proposed method leverages both spatial and frequency-domain features, applying robust descriptors such as Discrete Cosine Transform (DCT) and Scale-Invariant Feature Transform (SIFT) to identify duplicated regions regardless of rotation, scaling, or postprocessing. Advanced machine learning and deep learning approaches are also explored to enhance detection accuracy and reduce false positives in complex image backgrounds. Experimental evaluations conducted on benchmark image forgery datasets demonstrate that the proposed system achieves high precision and recall in identifying forged regions, even under challenging conditions such as noise addition, compression, and geometric transformations. By combining classical forensic techniques with modern intelligent approaches, Forensic Vision contributes to the growing field of digital forensics, ensuring the reliability of visual evidence in domains such as journalism, law enforcement, and cybersecurity.