| Network Reduction Using Error Prediction |
DEC 90 |
|
| Authors:
Michael V. Gilsdorf; AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF ENGINEERING
|
 | This thesis investigates gradient descent learning algorithms for multi-layer feed forward neural networks. A technique is developed which uses error prediction to reduce the number of weights/nodes in a network. The research begins by studying the first and second order back-prop training algorithms along with their convergence properties. A network is reduced by making an estimate of the amount of error which would occur when a weight(s) is removed. This ... |
|
| Image Acquisition and Processing Equipment for Machine Vision |
30 SEP 90 |
24 pages |
| Authors:
Donald E. McClure; BROWN UNIV PROVIDENCE RI
|
 | The Grant titled Image Acquisition and Processing Equipment for Machine Vision funded under the Defense University Research Instrumentation Program was used to obtain a sensor and state-of-the-art special-purpose computer equipment for research on problem sof image analysis and computer vision. Our approach to problems in image analysis is strongly rooted in a mathematical statistical foundation. The foundation builds from Pattern Theory and adopts a Bayesian formulation of image inference problems. ... |
|
| Non Parametric Classification Using Learning Vector Quantization |
21 AUG 90 |
|
| Authors:
Anthony LaVigna; MARYLAND UNIV COLLEGE PARK
|
| We study several properties of Learning Vector Quantization (LVQ). LVQ is a nonparametric detection scheme proposed in the neural network community by Kohonen. We examine it in detail, both theoretical and experimentally, to determine its properties as a nonparametric classifier. In particular, we study the convergence of the parameter adjustment rule in LVQ, we present a modification to LVQ which results in improving he convergence of the algorithm, we show ... |
|
| Signal Processing and Recognition in Adaptive Neural Networks |
09 AUG 90 |
|
| Authors:
Shihab Shamma; P. S. Krishnaprasad; MARYLAND UNIV COLLEGE PARK
|
| This research can be subdivided into four areas: (1) Models and neurophysiology of the auditory cortex. This includes mappings of physiological responses to sound, psychoacoustical studies, and mathematical models of the data. (2) Implementations of the cochlear and other auditory models both in DSP and VLSI forms. (3) Unsupervised learning algorithms applied to problems in sound segmentation, timbre characterization, and pitch extraction. (4) Applications of wavelet transforms to the analysis ... |
|
| SuperTroop via I-Port: Distributed Simulation Technology for Combat Development and Training Development |
AUG 90 |
|
| Authors:
Paul F. Gorman; INSTITUTE FOR DEFENSE ANALYSES ALEXANDRIA VA
|
| This paper proposes research and development aimed at total encapsulation of an individual who fights on foot, predicated upon integrating a powered exoskeleton into his battle dress to augment load-bearing capability, a personal computer networked with those of fellow combatants, and full body protection against ballistic, chemical, thermal, and directed energy threats. Fielded first would be a simulation of the eventual battle dress--termed ST, for SuperTroop--which could give individual combatants ... |
|
| Efficient Image Generation Using Localized Frequency Components Matched to Human Vision |
JUL 90 |
|
| Authors:
George A. Geri; Yehoshua Y. Zeevi; Moshe Porat; DAYTON UNIV OH RESEARCH INST
|
| Following a brief tutorial in the general area of image analysis, a formalism is presented for using the generalized Gabor approach to image representation in the combined frequency-position space. This approach uses elementary functions to which the human visual system is particularly sensitive and which are efficient for the analysis and synthesis of visual imagery. Among the topics covered are the complementarity of position and spatial frequency in the Gabor ... |
|
| Sizing-Up Molecular Electronic Devices |
JUL 90 |
|
| Authors:
Ann E. Tate; Jennifer L. Sloop; NAVAL SURFACE WARFARE CENTER DAHLGREN VA
|
| Biotechnology is of interest to the Navy in relation to organic or molecular computing. Molecular computing is a novel technology that offers promise to meet the requirements of future ship systems that must possess sophisticated information processing capabilities that can discern genuine threats from nonthreats or decoys and do this instantaneously. A molecular computer is an information processing system that uses organic or biomolecules to sense, transform, and output signals ... |
|
| Statistical Decisions Utilizing Neural Nets |
30 MAY 90 |
|
| Authors:
Jack Manata; George Schlenker; ARMY ARMAMENT MUNITIONS AND CHEMICAL COMMAND ROCK ISLAND IL SYSTEMS ANALYSIS OFFICE
|
| Neural networks were developed that accurately determine the statistical characteristics: modality and number fo stochastic components of underlying probability distribution(s) for sample data. Sample data examples, used to teach the neural nets were generated utilizing either a single beta distribution or a mixture of beta distributions. Once the neural net learned to distinguish between unimodal and multimodal examples and also between unimodal and mixture densities, they were challenged with unknown ... |
|
| Theoretical and Experimental Research Into Biological Mechanisms Underlying Learning and Memory |
24 APR 90 |
|
| Authors:
Leon N. Cooper; BROWN UNIV PROVIDENCE RI DEPT OF PHYSICS
|
| We describe an extended model of backward propagation incorporating gain modification and compare the performance of the extended model with ordinary backward propagation. We also describe our work on a statistical model for feature extraction based on the BCM neural network model. The model is presented as an exploratory (PP) (Projection Pursuit) algorithm. The formulation, which is similar in nature to PP, is based on a minimization of a cost ... |
|
| A Programming Environment for Parallel Vision Algorithms |
11 APR 90 |
|
| Authors:
Christopher Brown; ROCHESTER UNIV NY DEPT OF COMPUTER SCIENCE
|
| This contract developed and disseminated papers, ideas, algorithms, analysis, software, applications, and implementations for parallel programming environments for computer vision and for vision applications. The work has been widely reported and highly influential. The most significant work centered on the Butterfly Parallel Processor, the MaxVideo pipelined parallel image processor, and the development of the real-time computer vision laboratory. For the Butterfly, the Psyche multi-model operating system was developed and the ... |
|
| Field Measurement of Head Related Transfer Functions |
APR 90 |
|
| Authors:
Frederic Wightman; Doris J. Kistler; WISCONSIN UNIV-MADISON
|
| This effort sought to refine and simplify techniques for generating acoustic signals that could be used in three-dimensional auditory displays. Such signals are presented to a listener over headphones and create the illusion of a virtual sound source at a predetermined position in 3-D space. The signals are generated digitally, using algorithms based on the acoustic effects of human outer ear structures on sound waves reaching the ears. To date, ... |
|
| Collective Computation of Neural Network |
15 MAR 90 |
|
| Authors:
Guozheng Yao; Yunjiu Wang; FOREIGN TECHNOLOGY DIV WRIGHT-PATTERSON AFB OH
|
| Computational neuroscience is a new branch of neuroscience originating from current research on the theory of computer vision. The goal of the branch is to find a new interdisciplinary approach to dealing with information processing in the brain. The research involves scientists working in artificial intelligence engineering and neuroscience. The paper introduces the collective computational properties of model neural networks, mainly in a theoretical framework of the study of neural ... |
|
| Adaptive Control of Visually Guided Grasping in Neural Networks |
12 MAR 90 |
|
| Authors:
Michael Kuperstein; NEUROGEN LABS INC BROOKLINE MA
|
| We present a theory and prototype of a neural controller called INFANT that learns sensory-motor coordination from its own experience. INFANT adapts to unforseen changes in the geometry of the physical motor system and to the location, orientation, shape and size of objects. It can learn to accurately grasp an enlongated object without any information about the geometry of the physical sensory-motor system. This new neural controller relies on the ... |
|
| Non-Linear Analysis of Visual Cortical Neurons |
12 MAR 90 |
|
| Authors:
Lowell D. Jacobson; James P. Gaska; Daniel A. Pollen; MASSACHUSETTS UNIV MEDICAL SCHOOL WORCESTER
|
| During the first reporting period, new equipment was purchased and set up and new software was developed in preparation for electrophysiological experiments to study the neural networks that underly the binocular non-linear filtering properties of cells in the monkey (Macaca fascicularis) visual cortex. This preparatory task was completed. In addition, new methods were developed for using input-output measurements to identify multi-input nonlinear systems. These new mathematical results have been written ... |
|
| Joint Services Electronics Program |
31 JAN 90 |
|
| Authors:
W. G. Oldham; CALIFORNIA UNIV BERKELEY ELECTRONICS RESEARCH LAB
|
| The research is organized under two theme areas: High-speed wide-band elements for high-frequency electronics and new architectures for parallel computation. The former has 5 work units and one supplement ranging from optoelectronic, Gunn-effect, ultra-short-channel Si and superconductive devices to new techniques for electromagnetic scattering computation. The second theme area has 3 work units which cover research projects ranging from CAD techniques for the synthesis of very large circuits and investigations ... |
|
| Effect of Extraneous Color-Coded Targets on Identification of Targets on CRT Displays |
16 JAN 90 |
|
| Authors:
Saul M. Luria; David F. Neri; Matthew J. Shim; Robyn Bivenour; NAVAL SUBMARINE MEDICAL RESEARCH LAB GROTON CT
|
| Sonar operations, using the broadband waterfall display, search for a thin line which constitutes a target track. Although current displays are monochromatic, color may well be added in the future. To examine the usefulness of color coding the target tracks on such displays, we have measured the response times of individuals to lines coded with different colors; compared different sets of colors; measured the effect of varying the background colors; ... |
|
| An Introduction to the Information Processing Components of the Brain |
12 JAN 90 |
|
| Authors:
S. Collins; ROYAL SIGNALS AND RADAR ESTABLISHMENT MALVERN (UNITED KINGDOM)
|
| Over the past decade there has been increasing interest in neurological inspired computational techniques. This interest arises from the concurrence of two factors; firstly, a growing list of interesting tasks for which serial digital computers are unsuitable, and secondly, information gained from the application of new techniques in neurobiology. This text is intended to provide an introduction to neurobiological terms of physical scientists and engineers, with some pointers to further ... |
|
| Robotic Mobility System: Road Following Algorithms |
JAN 90 |
132 pages |
| Authors:
Robert Franklin; Dwayne C. Carmer; F. Pont; N. Finzel; L. Harmon; ENVIRONMENTAL RESEARCH INST OF MICHIGAN ANN ARBOR SENSOR SYSTEMS DIV
|
| Image processing algorithms were developed to permit use of a laser- based vision device for autonomous navigation of a wheeled vehicle. Imagery and processing results are given. Operation of the laser-based vision device (ERIM's Autonomous Land Vehicle sensor) is described, including a new output signal that is sensitive to scene texture or vertical structure; other signals provide range and reflectance information. Simulation software packages were developed to model a robotic ... |
|
| Simulating Biological Vision with Hybrid Neural Networks |
90 |
23 pages |
| Authors:
Paul Sajda; Leif H. Finkel; PENNSYLVANIA UNIV PHILADELPHIA
|
| We present an example of how vision systems can be modeled and designed by integrating a top-down computationally-based approach with a bottom- up biologically-motivated architecture. The specific visual processing task we address is occlusion-based object segmentation-the discrimination of objects using cues derived from object interposition. We construct a model of object segmentation using hybrid neural networks-distributed parallel systems consisting of neural units modeled at different levels of abstraction. We show ... |
|
| A General Architecture for Classical Conditioning of Perceptual-Motor Sequences |
1990 |
7 pages |
| Authors:
Michael R. Blackburn; NAVAL OCEAN SYSTEMS CENTER SAN DIEGO CA
|
| A general neural network architecture, loosely modeled on the cerebral cortex, for the classical conditioning of perceptual-motor sequences is described. The utility of such an algorithm in robotics applications lies in its potential to adaptively order gross and fine motor actions under sensory control. |
|
| An Intelligent Job Dispatcher for Computer Systems |
22 DEC 89 |
|
| Authors:
Jurn S. Leung; GENERAL PURPOSE MACHINES LAB IRVINE CA
|
| The Intelligent Job Dispatcher has two networks, the LMN-net and the information gathering network. The main result of this effort is the architecture of the LMN-net, which optimizes the assignment of N jobs to M processors in a window of L time steps. Input to this network are information on job loads, processor capacities, and inter-job dependence. Output of this network are assignment matrices. This network is a modified Hopfield ... |
|
| Visual Servoing for Autonomous Aircraft Refueling |
14 DEC 89 |
|
| Authors:
Richard P. Shipman; AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF ENGINEERING
|
| The design and development of autonomous robotic systems for autonomous aircraft refueling is a major interest to the Air Force. The approach being undertaken is the integration of vision and robot technologies allowing visual information, regarding identification and location of an aerial refueling receptacle on an aircraft, to be fed back to a robot controller which develops trajectory information needed for compliant robot motion. The purpose of this thesis effort ... |
|
| Gain Modification Enhances High Momentum Backward Propagation |
08 DEC 89 |
|
| Authors:
Charles M. Bachmann; BROWN UNIV PROVIDENCE RI CENTER FOR NEURAL SCIENCE
|
| We present a backward propagation network which simultaneously modifies the gain parameters and the synaptic weights. Gain modification is shown to enhance the improvement in convergence rate obtained by high momentum in standard synaptic backward propagation. These improvements occur without degrading the generalization capabilities of the final solutions obtained by the network. Keywords: Neural nets; Backward propagation; Gain modification; Momentum; Effective time-dependent; Step constant. |
|
| The Perceptual Interaction of Graphical Attributes in Thirteen Bivariate Displays |
DEC 89 |
|
| Authors:
C. M. Carswell; Christopher D. Wickens; LOUISVILLE UNIV KY
|
| Thirteen bivariate graphical displays were used to perform the speeded classification tasks necessary to diagnose dimensional integrality configurality, and separability. The nine classification tasks were performed by a different group of nine subjects for each of the graphs. Results from 9 of the 13 graphs clearly fit the pattern indicative of integrality. Instead, graphs varied on a continuum from highly configural to separable displays. A principal components analysis was used ... |
|
| Applications of Neural Networks to Adaptive Control |
DEC 89 |
|
| Authors:
Russell W. Scott Ii; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
| The amount of a priori knowledge required to design some modern control systems is becoming prohibitive. Two current methods addressing this problem are robust control, in which the control design is insensitive to errors in system knowledge, and adaptive control, in which the control law is adjusted in response to a continually updated model of the system. This thesis examines the application of parallel distributed processing (neural networks) to the ... |
|
| A Plant Bioassay for Assessing Plant Uptake of Heavy Metals from Contaminated Freshwater Dredged Material |
DEC 1989 |
9 pages |
| Authors:
Bobby L. Folsom Jr; Richard A. Price; Robert M. Engler; ARMY ENGINEER WATERWAYS EXPERIMENT STATION VICKSBURG MS ENVIRONMENTAL LAB
|
| The Decisionmaking Framework (DMF) developed by Peddicord et al. (1986) provides a framework for evaluating sediments before dredging. This framework is comprised of several modules one of which is the Plant Bioassay for materials proposed for upland or wetland placement. The purpose of this note Is to describe the methods and materials necessary to conduct such a plant bioassay. ... |
|
| Gain Modification Enhances High Momentum Backward Propagation |
30 NOV 89 |
|
| Authors:
Charles M. Bachmann; BROWN UNIV PROVIDENCE RI CENTER FOR NEURAL SCIENCE
|
| A backward propagation network is presented which simultaneously modifies the gain parameters and the synaptic weights. Gain modification is shown to enhance the improvement in convergence rate obtained by high momentum in standard synaptic backward propagation. These improvements occur without degrading the generalization capabilities of the final solutions obtained by the network. Keywords: Backward propagation; Gain modification; Momentum. (JHD) |
|
| Classification and Simulation of Single and Multi-Degree-of-Freedom Motions |
30 OCT 89 |
|
| Authors:
Bahram Ravani; CALIFORNIA UNIV DAVIS DEPT OF MECHANICAL ENGINEERING
|
| This report provides an outline of the results obtained in a period of one year in Simulation and Dynamic Motions (trajectories) of robot manipulators performing multi-degree-of-freedom motions. The design of a general purpose world modeling system for Computer Aided Design based robot motion simulation was completed. Methods were also developed for calibration of the world model to account for simple inaccuracies developed in the actual robot task operations. Methods were ... |
|
| Vision in Dynamic Environments |
15 AUG 89 |
|
| Authors:
Azriel Rosenfeld; MARYLAND UNIV COLLEGE PARK CENTER FOR AUTOMATION RESEARCH
|
| Research conducted on the contract was primarily concerned with real- time, three-dimensional computer vision and image understanding. Abstracts: Tracing Finite Motions without Correspondence. Group Theoretical Methods in Image Understanding; Interpreting Aerial Photographs by Segmentation and Search; Straight Line Fitting in a Noisy Image; Calibration of a Stereo System with Small Relative Angles; Cylindrical Shape from Contour and Shading without Knowledge of Lighting Conditions or Surface Albedo; A Robust Algorithm for ... |
|
| IEEE Conference on Neural Information Processing Systems - Natural and Synthetic Held in Denver, Colorado on 28 November-1 December 1988 |
14 AUG 89 |
|
| Authors:
Terrence Sejnowski; JOHNS HOPKINS UNIV BALTIMORE MD DEPT OF BIOPHYSICS
|
| Partial Contents: Speedy Alternatives to Back Propagation; Mean Field Annealing and Neural Networks; Comparison of Two LP Parametric Representations in a Neural Network-Based Speech Recognizer; Neural Net Receivers in Spread- Spectrum Multiple-Access Communication Systems; Modeling a Central Pattern Generator in Software and Hardware: Tritonia in Sea Moss; Storage of Covariance by the Selective Long-Term Potentiation and Depression of Synaptic Strengths in the Hippocampus; Convergence and Pattern-Stabilization in the Boltzmann Machine; ... |
|
| Parallel Vision Algorithm Design and Implementation 1988 End of Year Report |
AUG 89 |
|
| Authors:
Takeo Kanade; Jon A. Webb; CARNEGIE-MELLON UNIV PITTSBURGH PA ROBOTICS INST
|
| The Apply programming language has been extended to allow variable- sized image computations, and also to allow border mirroring, in which pixels accessed outside the borders are produced by copying pixels from the interior of the image. Implementation and design decisions are discussed. Apply and the Warp programming language W2 were used to implement the second DARPA image understanding benchmark. The results of this implementation are reported. Experience with this ... |
|
| A Genetic Adaptive System for Image Understanding and Learning Research. Phase 1 |
AUG 89 |
|
| Authors:
Dean Z. Douthat; Kevin W. Ross; INSTITUTE FOR THE STUDY OF INTELLIGENT SYSTEMS ANN ARBOR MI
|
| This report documents the efforts and results of initial phase research on machine learning directed at application for real time machine vision and automatic target recognition. The particular paradigm pursued is based on genetic algorithms and classifiers modeled on the summation of Mendelian genetic recombination, Darwinian selection and ecological notions of competition. This machine learning approach is strongly supported by sound statistical theory. A second thread of research was the ... |
|
| Design and Implementation of a Flexible Robot |
AUG 89 |
|
| Authors:
Andrew D. Christian; MASSACHUSETTS INST OF TECH CAMBRIDGE ARTIFICIAL INTELLIGENCE LAB
|
| This report describes the design, implementation, and control of a robot that has low natural frequencies of vibration. Insights into the problems of designing joint and link flexibility are discussed. The deflection of the robot under gravity is correlated with the fundamental frequency of vibration. Different link geometries and materials are evaluated. The robot has three rotary actuators and two links, with two actuators at the base and the third ... |
|
| A Colony Architecture for an Artificial Creature |
AUG 89 |
|
| Authors:
Jonathan Connell; MASSACHUSETTS INST OF TECH CAMBRIDGE ARTIFICIAL INTELLIGENCE LAB
|
| In this report we describe a distributed control system for a mobile robot which operates in an unmodified office environment occupied by moving people. The robot's controller is composed of over 40 separate processes which run on a loosely connected network of 24 processors. Together this ensemble helps the robot locate empty soda cans, collect them with its arm, and bring them back home. A multi-agent system such as this ... |
|
| Dynamic Image Interpretation For Autonomous Vehicle Navigation |
AUG 89 |
60 pages |
| Authors:
Edward M. Riseman; Allen R. Hanson; MASSACHUSETTS UNIV AMHERST DEPT OF COMPUTER AND INFORMATION SCIENCE
|
| This report presents the results of the Dynamic Image Interpretation for the Autonomous Vehicle Navigation project for the time period 2/26/85 to 7/ 12/89. The purpose of the project is to develop algorithms and tools to enable a robotic ground vehicle to navigate autonomously through realistic landscapes. In this final annual report, we summarize our accomplishments in constructing robust algorithms to be used for vehicle navigation as well as tools ... |
|
| Maximum Entropy Deconvolver Circuit Based on Neural Net Principles. |
18 JUL 1989 |
|
| Authors:
Martin C. Peckerar; Christie R. Marrian; DEPARTMENT OF THE NAVY WASHINGTON DC
|
| Disclosed are two modifications of the Tank-Hopfield circuit, each of which enables the deconvolution of a signal in the presence of noise. In each embodiment, the Tank-Hopfield circuit is modified so that the equation for total circuit energy reduces to one term representing convolution and another information theoretic (or Shannon) entropy. Thus, in finding its global minimum energy state, each modified circuit inherently identifies an optimal estimate of a deconvoluted ... |
|
| IEEE (Institute of Electrical and Electronic Engineers) Conference on Neural Information Processing Systems - Natural and Synthetic Held in Denver, Colorado on November 8-12, 1987. Abstracts of Papers |
30 JUN 89 |
|
| Authors:
E. C. Posner; INFORMATION THEORY SOCIETY (IEEE) PASADENA CA
|
| Contents: Bifurcation analysis of a network model of a rabbit olfactory bulb with periodic attractors stored by a sequence learning algorithm; Model of the cerebellum as an array of adjustable pattern generators; Spontaneous and information-triggered segments of series of human brain electric field maps; Towards an organizing principle for perception: The role of Hebbian synapses in the emergence of feature-analyzing function; Synchronization in neural nets; New neural algorithms for associative ... |
|
| JPRS Report, Science & Technology, Europe |
05 JUN 1989 |
|
| Authors:
JOINT PUBLICATIONS RESEARCH SERVICE ARLINGTON VA
|
| This report contains information concerning Europe's science and technology. The following topics are addressed: West Europe; (1) aerospace and civil aviation, (2) factory automation and robotics, (3) microelectronics, (4) science and technology policy, (5) technology transfer, East Europe; (1) computers, (2) factory automation and robotics, (3) nuclear engineering, and (4) telecommunications. |
|
| Role of Retinocortical Processing in Spatial Vision |
JUN 89 |
|
| Authors:
Donald H. Kelly; SRI INTERNATIONAL MENLO PARK CA
|
| The inhomogeneous retinal filtering algorithms is incorporated into a more general model that includes conformal projection of the retinal filtered outputs into cortical input images, suitable for further processing, such as Gabor filtering. Our new cortical images seem to show much less loss of information relative to the retina. We no longer restore some of the dc (zero- frequency) component that is filtered out by the (Laplacian/Gaussian) retinal receptive-field model. ... |
|
| Resistive Fuses: Analog Hardware for Detecting Discontinuities in Early Vision |
JUN 89 |
|
| Authors:
John Harris; Christof Koch; Jin Luo; John Wyatt; MASSACHUSETTS INST OF TECH CAMBRIDGE MICROSYSTEMS RESEARCH CENTER
|
| The detection of discontinuities in motion, intensity, color, and depth is a well studied but difficult problem in computer vision. We discuss our 'resistive fuse' circuit--the first hardware circuit that explicitly implements either analog or binary line processes in a controlled fashion. We have successfully designed and tested an analog CMOS VLSI circuit that contains a 1-D resistive network of fuses implementing piece-wise smooth surface interpolation. The segmentation ability of ... |
|
| JPRS Report, Science & Technology, Europe |
22 MAY 1989 |
|
| Authors:
JOINT PUBLICATIONS RESEARCH SERVICE ARLINGTON VA
|
|
| Parallel and Deterministic Algorithms from MRFs (Markov Random Fields): Surface Reconstruction and Integration |
MAY 89 |
|
| Authors:
Davi Geiger; Federico Girosi; MASSACHUSETTS INST OF TECH LEXINGTON LINCOLN LAB
|
| In recent years many researchers have investigated the use of Markov random fields (MRFs) for computer vision. They can be applied for example in the output of the visual processes to reconstruct surfaces from sparse and noisy depth data, or to integrate early vision processes to label physical discontinuities. Drawbacks of MRFs models have been the computational complexity of the implementation and the difficulty in estimating the parameters of the ... |
|
| On Interpreting Stereo Disparity |
APR 89 |
|
| Authors:
Richard P. Wildes; MASSACHUSETTS INST OF TECH CAMBRIDGE ARTIFICIAL INTELLIGENCE LAB
|
| The problems under consideration center around the interpretation of binocular stereo disparity. In particular, the goal is to establish a set of mappings from stereo disparity to corresponding three dimensional scene geometry. Stereo disparity is represented as a vector field derived from differential projection of a three dimensional scene onto a pair of two- dimensional imaging surfaces. The resulting disparity field is analysed with the aid of mathematical tools from ... |
|
| United States Air Force Research Initiation Program for 1987. Volume 4. |
APR 1989 |
|
| Authors:
Rodney C. Darrah; UNIVERSAL ENERGY SYSTEMS INC DAYTON OH
|
| Contents: Development of Implantable Devices for Sustained Delivery of Volatile Hydrocarbons in Rats; In Situ Detection of Osteoprogenitor Cells in an Actively Growing Bone System; Trauma-Activated Periosteum Derived Osteogenic Cells: Response to Selected Growth Factors; Assessing the Attributes of Expert Judgement: Measuring Bias in Subjective Uncertainty Estimates; Mathematical Modeling; Learning Behavior of Adaptive Filters for Evoked Brain Potentials; The Rhetoric of Hypertext: An Examination of Document Database Concepts and The ... |
|
| DARPA Neural Network Study: October 1987 - February 1988 |
22 MAR 89 |
|
| Authors:
MASSACHUSETTS INST OF TECH LEXINGTON LINCOLN LAB
|
| The Neural Network Study, sponsored by the Tactical Technology Office of the U.S. Defense Advanced Research Projects Agency (DARPA/TTO), was conducted under the auspices of the Massachusetts Institute of Technology's Lincoln Laboratory (MIT/LL) from October, 1987 through February, 1988 with government, industry, and academic participants. The goal of the study, all of which were achieved, were: To identify potential applications for neural networks in Department of Defense (DoD) systems, To ... |
|
| Computer Implementation and Simulation of Some Neural Networks Used in Pattern Recognition and Classification |
MAR 89 |
|
| Authors:
Mohamed H. Khaidar; NAVAL POSTGRADUATE SCHOOL MONTEREY CA
|
| Searchers and scientists have been studying neural networks for many years hoping to achieve human-like performance in the fields of speech and pattern recognition and classification. This study will first make an introduction to the field of artificial neural networks, then describe some of the neural nets used in the pattern recognition and classification. A computer simulation program from an algorithmic approach for each one of these networks will be ... |
|
| Northeast Artificial Intelligence Consortium Annual Report for 1987. Volume 4. Research in Automated Photointerpretation |
MAR 89 |
|
| Authors:
James Modestino; SYRACUSE UNIV NY
|
| The Northeast Artificial Intelligence Consortium (NAIC) purpose is to conduct pertinent research in artificial intelligence and to perform activities ancillary to this research. This report describes progress that has been made in the third year of the existence of the NAIC on the technical research tasks undertaken at the member universities. The topics covered in general are: versatile expert system for equipment maintenance, distributed AI for communications system control, automatic ... |
|
| Northeast Artificial Intelligence Consortium Annual Report for 1987. Volume 7. Parallel, Structural, and Optimal Techniques in Vision |
MAR 89 |
|
| Authors:
Christopher M. Brown; SYRACUSE UNIV NY
|
| Investigation is made of various aspects of parallel computer vision, with the goal of building behaving, real-time systems that perform multi-sensory integration. A commitment is made to the idea that an intimate coupling of sensory and motor capabilities is a way to make progress on the vision problem. Behaving animals have such a coupling, and its benefits have been demonstrated analytically. A special-purpose parallel pipelined hardware was acquired for low-level ... |
|
| Surface Reflection: Physical and Geometrical Perspectives |
MAR 89 |
|
| Authors:
Shree K. Nayar; Katsushi Ikeuchi; Takeo Kanade; CARNEGIE-MELLON UNIV PITTSBURGH PA ROBOTICS INST
|
| Machine vision can greatly benefit from the development of accurate reflectance models. There are two approaches to the study of reflection: physical and geometrical optics. While geometrical models may be construed as mere approximations to physical models, they possess simpler mathematical forms that often render them more usable than physical models. However, geometrical models are applicable only when the wavelength of incident light is small compared to the dimensions of ... |
|
| The Image Understanding Architecture Project |
MAR 89 |
|
| Authors:
Charles C. Weems; Steven P. Levitan; Allen R. Hanson; Edward M. Riseman; David Shu; MASSACHUSETTS UNIV AMHERST DEPT OF COMPUTER AND INFORMATION SCIENCE
|
| The primary goal of the Image Understanding Architecture (IUA) project is to build a proof-of-concept prototype of a 1/64th slice of a next generation vision architecture, and develop the software support environment that will be needed to utilize the hardware. The second year of this program has focussed on extensions to the IUA software simulator programming environment, the development of library routines and demonstration software for the IUA, construction of ... |
|
Bionics
