Research On Multiple Information Hiding In Multimedia Data

With the growing popularity of broadband Internet and the rapid development ofdigital multimedia (image, video and audio, etc.) technology, information hiding asan effective tool for copyright protection and covert communication has drawn moreand more attention. Traditional information hiding schemes are usually designed toachieve only one goal, such as copyright protection or authentication. There are manyissues such as low hiding space utilization, poor scalability and small capacity. As aneffective method to solve these problems, multiple information hiding has advantagesin multimedia data tracking, data usage monitoring, multiple property management andmany other applications.This dissertation proposes corresponding solutions to the above issues in multi-media information hiding schemes after thorough investigation and detailed analysis ofexisting information hiding algorithms. The main contributions are as follows.For low space utilization and single function issues in traditional image water-marking algorithms, we propose a dual image watermarking scheme. While the robustwatermark is embedded into robust coefficients, a fragile watermark is embedded intosensitive coefficients to ensure the image integrity, which adds watermarking featuresand improves the space utilization.For low space utilization in traditional video watermarking algorithms, we presenta dual watermarking scheme to hide information both in key frames and non-criticalframes simultaneously, which greatly improves the space utilization. In key frames, inorder to meet the real-time demand of live encoding watermark, a new prediction modewatermarkingschemeforrobustwatermarkembeddinginkeyframeisproposed, whichhas a low computational cost and invokes no information loss. In non-critical frames,a fragile watermark is embedded to detect malicious tampering.For small capacity and poor scalability issues in audio covert communication, wepropose a hybrid covert channel to take advantage of both storage and timing featuresof voice data flow, which can be combined with any existing storage channel schemeto increase channel capacity and enhance scalability. Based on the analysis and processing of multimedia data characteristics, this dis-sertation solves the current low space utilization, poor scalability and small capacityissues using a multiple information hiding method.The proposed algorithms have been implemented and compared with existing schemes.The results show that our information hiding algorithms achieve a better performancein both functionality and efficiency.