A. James Clark School of Engineering
Permanent URI for this communityhttp://hdl.handle.net/1903/1654
The collections in this community comprise faculty research works, as well as graduate theses and dissertations.
Browse
2 results
Search Results
Item Intrinsically Embedded Signatures for Multimedia Forensics(2016) Al Hajj Ahmad, Adi Mohammad; Wu, Min; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This dissertation examines the use of signatures that are intrinsically embedded in media recordings for studies and applications in multimedia forensics. These near-invisible signatures are fingerprints that are captured unintentionally in a recording due to influences from the environment in which it was made and the recording device that was used to make it. We focus on two types of such signatures: the Electric Network Frequency (ENF) signal and the flicker signal. The ENF is the frequency of power distribution networks and has a nominal value of 50Hz or 60Hz. The ENF fluctuates around its nominal value due to load changes in the grid. It is particularly relevant to multimedia forensics because ENF variations captured intrinsically in a media recording reflect the time and location related properties of the respective area in which it was made. This has led to a number of applications in information forensics and security, such as time-of-recording authentication/estimation and ENF-based detection of tampering in a recording. The first part of this dissertation considers the extraction and detection of the ENF signal. We discuss our proposed spectrum combining approach for ENF estimation that exploits the presence of ENF traces at several harmonics within the same recording to produce more accurate and robust ENF signal estimates. We also explore possible factors that can promote or hinder the capture of ENF traces in recordings, which is important for a better understanding of the real-world applicability of ENF signals. Next, we discuss novel real-world ENF-based applications proposed through this dissertation research. We discuss using the embedded ENF signal to identify the region-of-recording of a media signal through a pattern analysis and learning framework that distinguishes between ENF signals coming from different power grids. We also discuss the use of the ENF traces embedded in a video to characterize the video camera that had originally produced the video, an application that was inspired by our work on flicker forensics. The last part of the dissertation considers the flicker signal and its use in forensics. We address problems in the entertainment industry pertaining to movie piracy related investigations, where a pirated movie is formed by camcording media content shown on an LCD screen. The flicker signature can be inherently created in such a scenario due to the interplay between the back-light of an LCD screen and the recording mechanism of the video camera. We build an analytic model of the flicker, relating it to inner parameters of the video camera and the screen producing the video. We then demonstrate that solely analyzing such a pirated video can lead to the identification of the video camera and the screen that produced the video, which can be used as corroborating evidence in piracy investigations.Item Temporal and Spatial Alignment of Multimedia Signals(2014) Su, Hui; Wu, Min; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)With the increasing availability of cameras and other mobile devices, digital images and videos are becoming ubiquitous. Research efforts have been made to develop technologies that utilize multiple pieces of multimedia information simultaneously. This dissertation focuses on the temporal and spatial alignment of multimedia signals, which is a fundamental problem that needs to be solved to enable such applications dealing with multiple pieces of multimedia data. The first part of the dissertation addresses the synchronization of multimedia signals. We propose a new modality for audio and video synchronization based on the electric network frequency (ENF) signal naturally embedded in multimedia recordings. Synchronization of audio and video is achieved by aligning the ENF signals. The proposed method offers a significant departure to tackling the audio/video synchronization problem from existing work, and a strong potential to address previously untractable scenarios. Estimation of the ENF signal from video is a challenging task. In order to address the problem of insufficient sampling rate of video, we propose to exploit the rolling shutter mechanism commonly adopted in CMOS camera sensors. Several techniques are designed to alleviate the distortions of motions and brightness changes in videos for ENF estimation. We also address several challenges that are unique to the synchronization of digitized analog audio recordings. Speed offset often occurs in digitized analog audio recordings due to the inconsistency in the tape's rolling speed. We show that the ENF signal captured by the original analog audio recording can be retained in the digitized version. The ENF signal is considered approximately as a single-tone signal and used as a reference to detect and correct speed offsets automatically. A complete multimedia application system often needs to jointly consider both temporal synchronization and spatial alignment. The last part of the dissertation examines the quality assessment of local image features for efficient and robust spatial alignment. We propose a scheme to evaluate the quality of SIFT features in terms of their robustness and discriminability. A quality score is assigned to every SIFT feature based on its contrast value, scale and descriptor, using a quality metric kernel that is obtained in a one-time training phase. Feature selection is performed by retaining features with high quality scores. The proposed approach is also applicable to other local image features, such as the Speeded Up Robust Features (SURF).