Image Watermarking Using LTP and DCT

Abstract

When an image enters into the digital world, the probability that the image will get tempered also increases. Hence there exists a higher threat to such content (image, audio, and video) to be copied, modified and distributed. In this thesis, LTP (Local Ternary Pattern) along with DCT (Discrete Cosine Transform) have been employed. By using the DCT technique on the JPEG image we are able to provide lossy compression to the image. In this way, the image is compressed such that size of the image reduced enough for storage of content, in this case, a watermark image. Once this has been done, LTP technique is applied to it. The use of LTP will help to embed a watermark even if it is of a heavier size. The original image will be obtained with lesser impacts to it. To show the effectiveness of LTP watermarking, certain attacks like cropping, noise, and compression are applied. As a result, the extracted watermark is tempered. DCT ensures robustness by compressing the image while LTP allows watermark embedding and extraction. DWT and LBP both provide compression and watermarking respectively. But, this paper has briefly compared DCT and DWT techniques to conclude as to why DCT has been chosen. Also, this paper shows reasons as to why LTP is chosen over LBP. Index Terms-Image Compression, Image Watermarking, Discrete Cosine Transform (DCT), Discrete Wavelet Transform (DWT). Local Ternary Pattern (LTP), Local Binary Pattern (LBP).