An Intelligent Steganographic Scheme Using Video Frame Neighborhoods

Information security based on steganographic methods is increasingly being introduced into various areas of human activity. However, existing methods are limited in many respects. A particularly important task is to increase the volume of hidden information using a fixed container size. This paper describes a method for steganographic security of information based on video. Using videos allows for an increase in the amount of embedded secret information. The method is based on dividing the video into an ordered sequence of video frames and analyzing groups of video frames to select pixels in each video file, into which secret bits are embedded. Video frame analysis is performed by determining the difference between adjacent video frames and forming an array of numbers that determines the magnitude of the difference for the code of each pixel. The use of threshold processing made it possible to identify pixels in which a large number of secret bits could be embedded inside their codes compared to the LSB algorithm. Based on the analysis of adjacent video frames and the application of threshold processing, templates are formed according to which secret information is embedded in the video. Traditional steganographic methods, such as Least Significant Bit (LSB) substitution, face challenges related to limited embedding capacity and vulnerability to common signal processing attacks. These drawbacks restrict their effectiveness in practical, high-security data hiding scenarios. To overcome these limitations, we propose an intelligent video steganography technique based on interframe pixel differences and adaptive thresholding. By identifying regions with significant temporal variation, the method selectively embeds multiple secret bits using a threshold-guided LSB approach. Additionally, a dynamic duplication mechanism across color channels is employed to improve redundancy and robustness without compromising visual quality. Experimental results show a notable increase in both embedding capacity and resistance to compression and noise, outperforming traditional LSB-based techniques. These advancements suggest strong potential for real-world applications in secure communication and digital media protection.