Multimedia And Security Team

Given a digital image, digital forensics aims to answer a number of forensic questions related to origin and authenticity of the image such as

• How was the image captured?
• Was it captured using a digital standalone camera, cell phone camera, digital scanner, or was it generated using computer graphics?
• If the image was captured using a camera, which brand or model of camera was used to capture the image?
• Has the image been tampered or manipulated after capture?
• Does it contain any hidden information or stego data?

YouTube and other web services alike have revolutionized content sharing and online social networking by providing an easy-to-use platform for users to post and share video. At the same time, content owners have raised serious concerns on unauthorized uploads of copyrighted movies and TV shows to these websites, as witnessed by high-profile lawsuits filed against YouTube and Google. In order to deter copyright violation and more importantly, to help keep online communities alive legally, "content fingerprinting" technologies are deployed to compute a short string of bits to capture unique characteristics of each video and use it determine whether an uploaded video belongs to a set of copyrighted content or not. Content fingerprints are also used by such applications as Shazam on iPhone to use recordings of short audio clips to identify the song and provide information about the artist, the album, and where to buy.

Our research focuses on developing a better understing of content fingerprinting systems through theoretical modeling and analysis and answer questions regarding the identification performance, scalability and security. Through our analysis, we have also derived guidelines for improving the performance of commonly used modules in fingerprinting.

Information retrieval over the encrypted domain is an emerging technology that aims at secure, efficient, and accurate document retrieval from an encrypted database without the need for decryption. This interdisciplinary research, combining cryptography, information retrieval, and multimedia signal processing, offers a promising direction for a more secure and privacy preserving approach for managing and retrieving sensitive information stored on today’s fast-growing network clouds.

With the development of the Internet and multimedia processing techniques, the protection of multimedia content has become increasingly important.  While cryptographic encryption is a powerful tool for access control and confidentiality protection, the protection usually terminates once the content is delivered and decrypted. The urgent need of the research effort addressing post-delivery protections come from both national security and commercial applications.

Digital Fingerprinting is an emerging technology to protect multimedia from unauthorized redistribution. It embeds a unique ID into each user's copy, which can be extracted to help identify culprits when an unauthorized leak is found. A powerful, cost-effective attack is the collusion attack from a group of users, where the users combine their copies of the same content but with different fingerprints to generate a new version. If designed improperly, the fingerprints can be attenuated or even removed by the collusion attack. Our research team at the University of Maryland has been taking an interdisciplinary approach to conducting research on digital fingerprinting for multimedia content protection.  Our research addresses a number of issues, including theory, design, attacks, and counter-attacks (particularly anti-collusion) for fingerprinting and tracing traitors.

Image hash functions find extensive applications in content authentication, database search, and watermarking. In this work, we develop a novel algorithm for generating an image hash based on Fourier transform features and controlled randomization. We formulate the robustness of image hashing as a hypothesis testing problem and evaluate the performance under various image processing modifications. We show that the hash function is resilient to content-preserving modifications, such as moderate geometric and filtering distortions. We introduce a new generic framework to study and evaluate the security of image hashing systems. Under this framework, we model the hash values as random variables and quantify its uncertainty in terms of the differential entropy. Using this proposed security framework, we analyze the security of the proposed schemes and several representative image hashing methods. We then examine the security versus robustness trade-off and show that the proposed methods can provide excellent security and robustness.

With the advancement of video-compression technology and the wide deployment of wireless networks, there is an increasing demand for wireless video communication services, and many design challenges remain to be overcome. In this article, we discuss how to dynamically allocate resources according to the changing environments and requirements, so as to improve the overall system performance and ensure individual quality of service (QoS). Specifically, we consider two aspects with regard to design issues: cross-layer design, which jointly optimizes resource utilization from the physical layer to the application layer, and multiuser diversity, which explores source and channel heterogeneity for different users. We study how to efficiently transmit multiple video streams, encoded by current and future video codecs, over resource-limited wireless networks such as 3G/4G cellular system and future wireless local/metropolitan area networks (WLANs/WMANs)

Figure: Illustration of better performing concealment scheme between GSB and OASI on the “Lena” image: (white blocks) OASI performs better; (black blocks) GSB performs better; (gray blocks) GSB and OASI do not have significant performance difference.

In an error-prone transmission environment, error concealment is an effective technique to reconstruct the damaged visual content. Due to large variations of image characteristics, different concealment approaches are necessary to accommodate the different nature of the lost image content. In this paper, we address this issue and propose using classification to integrate the state-of-the-art error concealment techniques. The proposed approach takes advantage of multiple concealment algorithms and adaptively selects the suitable algorithm for each damaged image area. With growing awareness that the design of sender and receiver systems should be jointly considered for efficient and reliable multimedia communications, we proposed a set of classification-based block concealment schemes, including receiver-side classification, sender-side attachment, and sender-side embedding. Our experimental results provide extensive performance comparisons and demonstrate that the proposed classification-based error concealment approaches outperform the conventional approaches.