TDLCI: An efficient scheme for tamper detection and localization in color images

In the contemporary multimedia world, image authentication has become incredibly important. According to an Ericsson report, the amount of data on the internet has multiplied ten times between 2010 and 2020. With the expansion of digital space, new tools for manipulating digital information keep on emerging. These tools are used maliciously to alter digital content. The protection of multimedia data in such a situation is of utmost importance, along with maintaining its integrity and authenticity. In this study, we have devloped and evaluated a new scheme for localization and detection of tamper based on pixel differencing in color images called TDLCI. The cover image is first segmented into non-overlapping blocks of 2 × 2 so that difference values can be calculated. Every possible dissimilarity value is placed within one of several bands. The range intervals are determined by considering the visible system's sensitivity to value fluctuations across a spectrum of smoothness and contrast.To implement the TDLCI scheme, we first separate the image into Red, Green, and Blue planes. Bits of watermark information are obtained by combining the most-significant-bits (MSBs) of each pixel in a 2 × 2 block with a chaotic sequence calculated by the logistic map and embedding the result in the least significant bits (LSBs) of the chosen pixels. During watermark extraction, the altered blocks are detected and located. The proposed method achieves an average PSNR and SSIM values of 43.27 dB and 0.9891, respectively which is better than state-of-the art. The average false positive rate (FPR) for random tampering has been found to be 4.598 which is better than state-of-the art, making TDLCI a better choice for tamper detection and localization in color images.