Associated Video Frame Steganography Model based on Frame Integrity Verification using GAN Network
Sameerunnisa Shaik, J. Jabez · 2024
Although advancements in Internet technology have made it easier to transmit and share information, they have also introduced new, more pervasive security concerns related to keeping that information safe. As discussed in this research, an upgraded steganography network was built to safeguard confidential image data and generate confidential and private information; this network comprises a hidden network and a revealing network to accomplish image embedding and recovery separately. The transmission of sensitive information via the public internet must always be protected. Protecting sensitive data from being stolen, accessed without authorisation, or lost in transit is a top priority in data security. Security of data transmitted may be guaranteed using variety of techniques, like as watermarking, cryptography, and steganography. In applications that operate in real time, growing bulk of video data necessitates strong security method to protect secret message while steganography decoding and encoding. Fundamental idea behind steganography is to encode data into file then transmit it covertly while keeping contents concealed. More data can be hidden using Generative Adversarial Networks GAN-based steganography than with more conventional approaches, and the technology remains undetectable. To generate fresh, synthetic occurrences of data that can pass for accurate data, GANs are computational frameworks that use two neural networks competing against each other. Such techniques differ in numerous characteristics: load capability, security, effectiveness, simplicity, and much more. The primary purpose of the process of Steganography is to hide large amounts of personal information against third-party attacks. Traditional models such as Advanced DCT and DWT LSB techniques disguise the critical data in each frame. These approaches, however, could be better suited to massive volumes of unstructured or compressed video data. Due to the high computational time and size requirements, most of these models rely on a plain secret message to embed and extract information from the short video datasets. To tighten security, it is possible to combine the ideas of encryption and Steganography. This study proposes an Associated Video Frame Steganography Model based on Frame Integrity Verification (AVFSM-FIV) on substantial real-time video datasets. As a means of ensuring the safety of the data embedded in experiments, the extraction time, accuracy, and vulnerability to attacks are simulated using several models of video steganography.