Computer Science Theses and Dissertations
Permanent URI for this collectionhttp://hdl.handle.net/1903/2756
Browse
2 results
Search Results
Item ROBUST TECHNIQUES FOR VISUAL SURVEILLANCE(2008-07-29) Tran, Son Dinh; David, Larry S; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The work described here aims at improving the performance of three building blocks of visual surveillance systems: foreground detection, object tracking and event detection. First, a new background subtraction algorithm is presented for foreground detection. The background model is built with a set of codewords for every pixel. The codeword contains the pixel's principle color and a tangent vector that represents the color variation at that pixel. As the scene illumination changes, a pixel's color is predicted using a linear model of the codeword and the codeword, in turn, is updated using the new observation. We carried out a number of experiments on sequences that have extensive lighting change and compare with previously developed algorithms. Second, we describe a multi-resolution tracking framework developed with efficiency and robustness in mind. Efficiency is achieved by processing low resolution data whenever possible. Robustness results from multiple level coarse-to-fine searching in the tracking state space. We combine sequential filtering both in time and resolution levels int a probabilistic framework. A color blob tracker is implemented and the tracking results are evaluated in a number of experiments. Third, we present a tracking algorithm based on motion analysis of regional affine invariant image features. The tracked object is represented with a probabilistic occupancy map. Using this map as support, regional features are detected and matched across frames. The motion of pixels is then established based on the feature motion. The object occupancy map is in turn updated according to the pixel motion consistency. We describe experiments to measure the sensitivity of our approach to inaccuracy in initialization, and compare it with other approaches. Fourth, we address the problem of visual event recognition in surveillance where noise and missing observations are serious problems. Common sense domain knowledge is exploited to overcome them. The knowledge is represented as first- order logic production rules with associated weights to indicate their confidence. These rules are used in combination with a relaxed deduction algorithm to construct a network of grounded atoms, the Markov Logic Network. The network is used to perform probabilistic inference for input queries about events of interest. The system's performance is demonstrated on a number of videos from a parking lot domain that contains complex interactions of people and vehicles.Item Algorithms and evaluation for object detection and tracking in computer vision(2005-08-01) Kim, Kyungnam; Davis, Larry; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Vision-based object detection and tracking, especially for video surveillance applications, is studied from algorithms to performance evaluation. This dissertation is composed of four topics: (1) Background Modeling and Detection, (2) Performance Evaluation of Sensitive Target Detection, (3) Multi-view Multi-target Multi-Hypothesis Segmentation and Tracking of People, and (4) A Fine-Structure Image/Video Quality Measure. First, we present a real-time algorithm for foreground-background segmentation. It allows us to capture structural background variation due to periodic-like motion over a long period of time under limited memory. Our codebook-based representation is efficient in memory and speed compared with other background modeling techniques. Our method can handle scenes containing moving backgrounds or illumination variations, and it achieves robust detection for different types of videos. In addition to the basic algorithm, three features improving the algorithm are presented - Automatic Parameter Estimation, Layered Modeling/Detection and Adaptive Codebook Updating. Second, we introduce a performance evaluation methodology called Perturbation Detection Rate (PDR) analysis for measuring performance of foreground-background segmentation. It does not require foreground targets or knowledge of foreground distributions. It measures the sensitivity of a background subtraction algorithm in detecting possible low contrast targets against the background as a function of contrast. We compare four background subtraction algorithms using the methodology. Third, a multi-view multi-hypothesis approach to segmenting and tracking multiple persons on a ground plane is proposed. The tracking state space is the set of ground points of the people being tracked. During tracking, several iterations of segmentation are performed using information from human appearance models and ground plane homography. Two innovations are made in this chapter - (1) To more precisely locate the ground location of a person, all center vertical axes of the person across views are mapped to the top-view plane to find the intersection point. (2) To tackle the explosive state space due to multiple targets and views, iterative segmentation-searching is incorporated into a particle filtering framework. By searching for people's ground point locations from segmentations, a set of a few good particles can be identified, resulting in low computational cost. In addition, even if all the particles are away from the true ground point, some of them move towards the true one through the iterated process as long as they are located nearby. Finally, an objective no-reference measure is presented to assess fine-structure image/video quality. The proposed measure using local statistics reflects image degradation well in terms of noise and blur.