In this thesis we present an effective steganalysis technique for digital video sequences based on the collusion attack. Steganalysis is the process of detecting with a high probability the presence of covert data in multimedia. Existing algorithms for steganalysis target detecting covert information in still images. When applied directly to video sequences these approaches are suboptimal. In this thesis we present methods that overcome this limitation by using redundant information present in the temporal domain to detect covert messages in the form of Gaussian watermarks. In particular we target the spread spectrum steganography method because of its widespread use. Our gains are achieved by exploiting the collusion attack that has recently been studied in the field of digital video watermarking and more sophisticated pattern recognition tools. Through analysis and simulations we, evaluate the effectiveness of the video steganalysis method based on averaging based collusion scheme. Other forms of collusion attack in the form of weighted linear collusion and block-based collusion schemes have been proposed to improve the detection performance. The proposed steganalsyis methods were successful in detecting hidden watermarks bearing low SNR with high accuracy. The simulation results also show the improved performance of the proposed temporal based methods over the spatial methods. We conclude that the essence of future video steganalysis techniques lies in the exploitation of the temporal redundancy.

An experimental approach to traffic flow analysis is presented in which methodology from pattern recognition is applied to a specific dataset to examine its utility in determining traffic patterns. The selected dataset for this work, taken from a 1985 study by JHK and Associates (traffic research) for the Federal Highway Administration, covers an hour long time period over a quarter mile section and includes nine different identifying features for traffic at any given time. The initial step is to select the most pertinent of these features as a target for extraction and local storage during the experiment. The tools created for this approach, a two-level hierarchical group of operators, are used to extract features from the dataset to create a feature space; this is done to minimize the experimental set to a matrix of desirable attributes from the vehicles on the roadway. The application is to identify if this data can be readily parsed into four distinct traffic states; in this case, the state of a vehicle is defined by its velocity and acceleration at a selected timestamp. A three-dimensional plot is used, with color as the third dimension and seen from a top-down perspective, to initially identify vehicle states in a section of roadway over a selected section of time. This is followed by applying k-means clustering, in this case with k=4 to match the four distinct traffic states, to the feature space to examine its viability in determining the states of vehicles in a time section. The method's accuracy is viewed through silhouette plots. Finally, a group of experiments run through a decision-tree architecture is compared to the kmeans clustering approach. Each decision-tree format uses sets of predefined values for velocity and acceleration to parse the data into the four states; modifications are made to acceleration and deceleration values to examine different results. The three-dimensional plots provide a visual example of congested traffic for use in performing visual comparisons of the clustering results. The silhouette plot results of the k-means experiments show inaccuracy for certain clusters; on the other hand, the decision-tree work shows promise for future work.

The primary objective of this thesis is to develop a methodology for clustering of objects based on their functionality typified by the notion of concept. We begin by giving a formal definition of concept. By assigning a functional interpretation to the underlying concept, we demonstrate the applicability of the functionally interpreted concept for clustering objects. This functional interpretation leads us to identifying two classes of concepts, namely, the Necessary class and the Quality-Improvement class. Next, we categorize the functional cohesiveness among objects into three different classes. Further, we axiomatize the restrictions imposed, on the execution of functions of objects, by the non-availability of sufficient resources. To facilitate describing functional clusters in a succinct manner, we define connectives that capture the imposed restrictions. Also we justify the adequacy of these connectives for describing functional clusters. We then propose a suitable data structure to represent the functionally interpreted concept, and develop an algorithm to perform this axiomatic functional partitioning of objects. We illustrate the functional partitioning of objects through a real-world example. We formally establish the invariance of the resulting cluster descriptions, with respect to the order in which the given set of objects is examined. This invariance would facilitate parallel implementations of the proposed methodology. We then analyze different functional cluster configurations from a structural viewpoint. In doing so, we identify the presence of a specific property among certain cluster configurations. We also state a sufficient condition for the presence of this property in any cluster. A separate class of concepts, namely the Concept Transformer class, displaying certain properties, is identified and studied in detail. We also demonstrate its applicability to functional clustering. Finally, we examine a knowledge-based pattern synthesis problem from a functional angle as a significant application of the functional interpretation of concept and associated data structures. Here, we show that a concept, from the functional view-point, can be viewed as the synthesis of various other concepts; the synthesis is an outcome of a knowledge-based goal-directed pattern-matching activity. The proposed methodology has the potential to cluster objects that imply functions by virtue of their physical properties.

Thesis (M.S.) -- New Jersey Institute of Technology, Dept. of Computer and Information Science, 2000. / Includes bibliographical references. Also available via the World Wide Web.

Thesis (M.S.)--University of Florida, 2001. / Title from first page of PDF file. Document formatted into pages; contains viii, 49 p.; also contains graphics. Vita. Includes bibliographical references (p. 47-48).

Thesis (Ph. D.)--State University of New York at Buffalo, 2001. / "April, 2001." Includes bibliographical references (leaves 195-211). Also available in print.

Thesis (Ph. D.)--University of Texas at Austin, 2003. / Vita. Includes bibliographical references. Available also from UMI Company.

Thesis (M. Sc.)--York University, 1999. Graduate Programme in Computer Science. / Typescript. Includes bibliographical references (leaves 90-93). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://wwwlib.umi.com/cr/yorku/fullcit?pMQ39223.

Thesis (M. Phil.)--Hong Kong University of Science and Technology, 2003. / Includes bibliographical references (leaves 60-63). Also available in electronic version. Access restricted to campus users.

Thesis (Ph. D.)--Carnegie Mellon University, 2003. / "October 2003." Includes bibliographical references.