Abstract
The broad use of high-performance tools for image acquisition and strong image processing software has made it easy for malicious purposes to manipulate images. Image splicing, which posed a threat to image integrity and authenticity, is a very popular and easy image forgery trick. Therefore, detection of image splicing is one of the major problems in digital forensics. A new passive (non-intrusive) image tampering detection technique is proposed here to detect splicing forgery based on discrete wavelet transform (DWT) and local binary pattern (LBP). First, input image is converted into YCbCr channels, and then, chroma channels are used as input image for feature extraction using 5-bin histogram and 3-CF moments from DWT domain. Then, ensemble classifier is used for detection of spliced and authentic images.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
