Institute for Systems Research Technical Reports
Permanent URI for this collectionhttp://hdl.handle.net/1903/4376
This archive contains a collection of reports generated by the faculty and students of the Institute for Systems Research (ISR), a permanent, interdisciplinary research unit in the A. James Clark School of Engineering at the University of Maryland. ISR-based projects are conducted through partnerships with industry and government, bringing together faculty and students from multiple academic departments and colleges across the university.
Browse
12 results
Search Results
Item Scalable Coding of Video Objects(1998) Haridasan, Radhakrishan; Baras, John S.; Baras, John S.; ISR; CSHCNThis paper provides a methodology to encode video objects in a scalable manner with regard to both content and quality. Content scalability and quality scalability have been identified as required features in order to support video coding across different environments. Following the object-based approach to coding video, we extend our previous work on motion-based segmentation by using a time recursive approach to segmenting image sequences and decomposing a video "shot" into its constituent objects. Our formulation of the segmentation problem enables us to design a codec in which the information (shape, texture and motion) pertaining to each video object is encoded independently of the other. The multiresolution wavelet decomposition used in encoding texture information is shown to be helpful in providing spatial scalability. Our codec design is also shown to be temporally scalable. This report was accepted for oral presentation at the IEEE International Symposium on Circuits & Systems, Monterey, Calif., May-June 1998.Item Combined Compression and Classification with Learning Vector Quantization(1998) Baras, John S.; Dey, Subhrakanti; ISRCombined compression and classification problems are becoming increasinglyimportant in many applications with large amounts of sensory data andlarge sets of classes. These applications range from aided target recognition(ATR), to medicaldiagnosis, to speech recognition, to fault detection and identificationin manufacturing systems. In this paper, we develop and analyze a learningvector quantization-based (LVQ) algorithm for the combined compressionand classification problem. We show convergence of the algorithm usingtechniques from stochastic approximation, namely, the ODE method. Weillustrate the performance of our algorithm with some examples.Item Accurate Segmentation and Estimation of Parametric Motion Fields for Object-based Video Coding using Mean Field Theory(1997) Haridasan, Radhakrishan; Baras, John S.; ISR; CSHCNWe formulate the problem of decomposing a scene into its constituent objects as one of partitioning the current frame into objects comprising it. The motion parameter is modeled as a nonrandom but unknown quantity and the problem is posed as one of Maximum Likelihood (ML) estimation. The MRF potentials which characterize the underlying segmentation field are defined in a way that the spatio-temporal segmentation is constrained by the static image segmentation of the current frame. To compute the motion parameter vector and the segmentation simultaneously we use the Expectation Maximization (EM) algorithm. The E-step of the EM algorithm, which computes the conditional expectation of the segmentation field, now reflects interdependencies more accurately because of neighborhood interactions. We take recourse to Mean Field theory to compute the expected value of the conditional MRF. Robust M-estimation methods are used in the M- step. To allow for motions of large magnitudes image frames are represented at various scales and the EM procedure is embedded in a hierarchical coarse-to-fine framework. Our formulation results in a highly parallel algorithm that computes robust and accurate segmentations as well as motion vectors for use in low bit rate video coding.This report has been submitted as a paper to the SPIE conference on Visual Communications and Image Processing - VCIP98 to be held in San Jose, California on Jan 24- 30, 1998. Item Estimation of Hidden Markov Models for Partially Observed Risk Sensitive Control Problems(1997) Frankpitt, Bernard A.; Baras, John S.; ISRWe look at the problem of estimation for partially observed, risk-sensitive control problems with finite state, input and output sets, and receding horizon. We describe architectures for risk sensitive controllers, and estimation, and we state conditions under which both the estimated model converges to the true model, and the control policy will converge to the optimal risk sensitive policy.Item Existence and Construction of Optimal Wavelet Basis for Signal Representation(1994) Zhuang, Y.; Baras, John S.; ISR; CSHCNWe study the problem of choosing the optimal wavelet basis with compact support for signal representation and provide a general algorithm for computing the optimal wavelet basis. We first briefly review the multiresolution property of wavelet decomposition and the conditions for generating a basis of compactly supported discrete wavelets in terms of properties of quadrature mirror filter (QMF) banks. We then parametrize the mother wavelet and scaling function through a set of real coefficients. We further introduce the concept of decomposition entropy as an information measure to describe the distance between the given signal and its projection onto the subspace spanned by the wavelet basis in which the signal is to be reconstructed. The optimal basis for a given signal is obtained through minimizing this information measure. We have obtained explicitly the sensitivity of dilations and shifts of the mother wavelet with respect to the coefficient set. A systematic approach is developed in this paper to derive the information gradient with respect to the parameter set from a given square integrable signal and a discrete basis of wavelets. The existence of the optimal basis for the wavelets has been proven in this paper. a gradient based optimization algorithm is developed for computing the optimal wavelet basis.Item Optimal Wavelet Basis Selection for Signal Representation(1994) Zhuang, Y.; Baras, John S.; ISR; CSHCNWe study the problem of choosing the optimal wavelet basis with compact support for signal representation and provide a general algorithm for computing the optimal wavelet basis. We first briefly review the multiresolution property of wavelet decomposition and the conditions for generating a basis of compactly supported discrete wavelets in terms of properties of quadrature mirror filter (QMF) banks. We then parametrize the mother wavelet and scaling function through a set of real coefficients. We further introduce the concept of information measure as a distance measure between the signal and its projection onto the subspace spanned by the wavelet basis in which the signal is to be reconstructed. The optimal basis for a given signal is obtained through minimizing this information measure. We have obtained explicitly the sensitivity of dilations and shifts of the mother wavelet with respect to the coefficient set. A systematic approach is developed here to derive the information gradient with respect to the parameter set for a given square integrable signal and the optimal wavelet basis. A gradient based optimazation algorithm is developed in this paper for computing the optimal wavelet basis.Item Identification of Infinite Dimensional Systems Via Adaptive Wavelet Neural Networks(1993) Zhuang, Y.; Baras, John S.; ISRWe consider identification of distributed systems via adaptive wavelet neural networks (AWNNs). We take advantage of the multiresolution property of wavelet systems and the computational structure of neural networks to approximate the unknown plant successively. A systematic approach is developed in this paper to find the optimal discrete orthonormal wavelet basis with compact support for spanning the subspaces employed for system identification. We then apply backpropagation algorithm to train the network with supervision to emulate the unknown system. This work is applicable to signal representation and compression under the optimal orthonormal wavelet basis in addition to autoregressive system identification and modeling. We anticipate that this work be intuitive for practical applications in the areas of controls and signal processing.Item Time-Recursive Computation and Real-Time Parallel Architectures, Part I: Framework(1993) Frantzeskakis, Emmanuel N.; Baras, John S.; Liu, K.J. Ray; ISRThe time-recursive computation has been proved as a particularly useful tool in real-time data compression, in transform domain adaptive filtering and in spectrum analysis. Unlike the FFT based ones, the time-recursive architectures require only local communication. Also, they are modular and regular, thus they are very appropriate for VLSI implementation and they allow high degree of parallelism. In this two part paper, we establish an architectural frame work for parallel time-recursive computation. In part I, we consider a class of linear operators that consists of the discrete time, time invariant, compactly supported, but otherwise arbitrary kernel functions. We show that the structure of the realization of a given linear operator is dictated by the decomposition of the latter with respect to proper basis functions. An optimal way for carrying out this decomposition is demonstrated. The parametric forms of the basis functions are identified and their properties pertinent to the architecture design are studied. A library of architectural building modules capable of realizing these functions is developed. An analysis of the implementation complexity for the aforementioned modules is conducted. Based on this framework, an architecture design procedure is developed in Part II [12] that can be used for routinely obtaining the time-recursive architecture of a given linear operator.Item A Theory of Adaptive Quasi Linear Representations(1992) Berman, Zeev; Baras, John S.; ISRThe analysis of the discrete multiscale edge representation is considered. A general signal description, called an inherently bounded Adaptive Quasi Linear Representation (AQLR), motivated by two important examples: the wavelet maxima representation and the wavelet zero-crossing representation, is introduced. This paper addresses the questions of uniqueness, stability, and reconstruction. It is shown, that the dyadic wavelet maxima (zero-crossings) representation is, in general, nonunique. Namely, for all maxima (zero-crossings) representation based on a dyadic wavelet transform, there exists a sequence having a nonunique representation. Nevertheless, these representations are always stable. Using the idea of the inherently bounded AQLR two stability results are proven. For a general perturbation, a global BIBO stability is shown. For a special case, where perturbations are limited to the continuous part of the representation, a Lipschitz condition is satisfied. A reconstruction algorithm, based on the minimization of an appropriate cost function, is proposed. The convergence of the algorithm is guaranteed for all inherently bounded AQLR. In the case, where the representation is based on a wavelet transform, this method yields an efficient, parallel algorithm, especially promising in an analog-hardware implementation.Item Structure of Divisible Discrete Random Sets and Their Randomized Superpositions(1991) Sidiropoulos, N.; Baras, John S.; Berenstein, Carlos A.; ISRIn this paper, we present an axiomatic formulation of Discrete Random Sets, and extend Choquet's uniqueness result to obtain a recursive procedure for the computation of the underlying event- space probability law, given a consistent Discrete Random Set specification via its generating functional. Based on this extension, we investigate the structure of Discrete Random Set models that enjoy the properties of independent decomposition/superposition, and present a design methodology for deriving models that are guaranteed to be consistent with some underlying event-space probability law. These results pave the way for the construction of various interesting models, and the solution of statistical inference problems for Discrete Random Sets.