No abstract

No abstract

The popularity of the internet has grown recently, making it simpler to broadcast and purchase digital content. This popularity suggests that counterfeiting has increased the susceptibility of multimedia to modification. A copy-move forgery is one of the most popular techniques for digital photo modification. Key point-based detection systems have successfully exposed copy-move forgery as a result of various digital assaults. On the other hand, these methods are useless when forgeries only involve tiny, smooth areas with a small number of crucial spots. A multi-scale feature matching based on the copy-move forgery detection technique was put into place to solve these problems. I believe the copy move method has some potential as well. Reduce the intensity threshold and rescale the supplied image to produce a reasonable number of key points in order to address those concerns in the proposed system initially. Then, a special multi-scale matching method is developed using scale clustering, overlapping grey scale clustering, and group matching modules. Present a novel iterative localization strategy that locates the forged region by utilising the resilience properties of each key points as well as its colour information. According to the experimental results, the suggested work's copy move forgery average is 1.375, compared to the existing system's average of 2.75, with standard deviations of 0.72 and 1.44, respectively. As a result, we can say that the proposed approach is more accurate and efficient than existing methods by more than 99.9%.

People are capable of manipulating and copying digital images. Protection and security must be given to the right property from piracy and illegal copies. To safeguard the intellectual property of digital images digital watermarking is used as a solution. Watermarking technique is prevalent due to an effective copyright protection method i.e., only the rightful owner can identify the signature that is embedded in a digital document. In this paper, a blind watermarking scheme is used which is dependent on the image content of certain specific DWT-DCT coefficients. The use of generalized Gaussian distributions helps in modelling the DCT-DWT coefficients of the Source image. The suggested system attained optimal PSNR, MSE and SSIM values and numerous attacks are resisted by an embedded watermark in the image watermarked as per the results of the experiment. Furthermore, the suggested system depicts far more robustness against numerous image processing attacks for example, scaling to 0.25 & 0.50, Cropping, and various types of noise attacks. Experimental results and numerical values of parameters confirms that the hybrid DCT-DWT blind watermarking technique is relatively efficient and also justifies the theoretical analysis.