Improving the Efficiency of Image and Video Forgery Detection Using Hybrid Convolutional Neural Networks

Abstract

Recently, on the internet, the level of image and video forgery has augmented due to the augmentation in the malware, which has facilitated user (anyone) to upload, download, or share objects online comprising audio, images, or video. Recently, Convolution Neural Network (CNN) has turn into a de-facto technique for classification of multi-dimensional data and it renders standard and also highly effectual network layer arrangements. But these architectures are limited by the speed due to massive number of calculations needed for training in addition to testing the network and also, it might render less accuracy. To trounce these issues, this paper proposed to ameliorate the image and video forgery detection’s efficiency utilizing hybrid CNN. Initially, the intensive along with incremental learning phase is carried out. After that, the hybrid CNN is implemented to detect the image together with video forgery. The developed system was tested on images together with videos for different kinds of forgeries, and it was observed that the proposed work obtains more than 98% accuracy for both testing as well as validation sets.

Survey on deep learning applications in digital image security

“…As can be seen in Table 3, when watermarked images face the attacks like median filtering of 3 × 3, salt and pepper noise with noise level δ of 0.005, and JPEG compression with quality factor 20, the deep learning method in Ref. 63 reaches almost all of the highest watermark quality assessment values, which are shown in bold, while it also obtains the most complete watermark with respect to no attack.…”

Management optimization of video system in the community

“…Therefore, in community work, it is necessary to have certain requirements for the video equipment and the image recorded by the equipment. The requirements for video equipment and video image quality have also brought many derived problems to the staff, such as how to deal with video color cast [5]. For example, in the work, it is necessary to effectively extract, analyze and judge the color and morphological features of suspicious vehicles in the video, and the physical features, clothing features and color features of suspicious people.…”