Carl V. Page

A Region-Based Approach to Digital Image Registration with Subpixel Accuracy

Automatic registration of images with translational, rotational, and scaling differences is discussed. To register two images from the same scene, first, the images are segmented and closedboundary regions in the images are extracted. Next, centers of gravity of closed-boundary regions are taken as control points and correspondence is established between the control points. Using this correspondence, the original images are then revisited and the segmentation process is refined in such a way that the obtained corresponding regions become optimally similar. This enables determination of centers of gravity of the regions up to subpixel accuracy. Finally, registration parameters are determined by the least squares error criterion.

Intermittent Faults: A Model and a Detection Procedure

Intermittent faults are those faults whose effects on the behavior of a system are present only part of the time. A probabilistic model for intermittent faults in digital circuits is suggested. A procedure for the detection of such faults in combinational circuits is proposed. The procedure employs the repeated application of tests that test for these faults as if their effects were permanent. The procedure is analogous to a sequential statistical decision problem. Least upper bounds on the number of repetitions of tests that detect a particular fault are derived. These bounds are then employed in designing optimum detection experiments. Such an optimization problem is found to be equivalent to an integer programming problem.

The use of Artificial Intelligence techniques in Computer-Assisted Instruction: an overview

One of the major goals of research in Artificial Intelligence is the representation of knowledge so that a computer can solve problems or communicate in a manner which exhibits “common sense”. Few programs for computers, including those for education, possess behavior which approaches any facet of the constellation of human skills and knowledge which are imprecisely called “common sense”. However, the revolutionary decline in hardware costs now makes it possible to consider economically viable, sophisticated designs for computer-aided instruction systems possessing some of the common sense attributes of a human tutor. In this survey we examine, in depth, techniques from Artificial Intelligence that can be used to endow a Computer-Aided Instruction system with approximations to some of the desirable qualities of a human tutor. We consider both techniques which have been proved in prototype systems for Computer-Aided Instruction and some techniques which were originally developed for other purposes.

Augmented Relaxation Labeling and Dynamic Relaxation Labeling

Current implementations of relaxation labeling are homogeneous, where each pixel is in an identical relationship to a static neighbor set. These systems maintain the iterative probabilistic labeling but use a nonhomogeneous dynamic neighborhood to establish a local consistency. Neighborhoods are created at each iteration through the broadcasting and reception of label information according to semantically established broadcasting patterns for each label. Augmented relaxation labeling is a two stage process which contains a separate relaxation stage with a top-down direction capability for specific pixel label updating. Dynamic relaxation is a one step process where every pixel label is updated through the dynamic neighborhoods. Both labeling processes are demonstrated on simple line drawings.

Parallel Contraction of Fibonacci Trees and Prefix Computations on a Family of Interconnection Topologies

The computation of prefixes of a given sequence (prefix computation [14]) and the fast reduction of a tree to a single node (tree contraction [1]) are two useful primitives for many applications on parallel computers. Most previous parallel algorithms have been based on the shared-memory model. We present general parallel algorithms for reducing a class of trees and prefix computations under the distributed-memory model. The new algorithms are shown to be communication-efficient and suitable for a large family of parallel computers. This family of parallel computers are based on a novel interconnection topology called the p-th order Fibonacci cube [10] that generalizes the Boolean cube (hypercube) [18] and the (second order) Fibonacci cube [8]. Specifically, the following results are presented: 1. We show that the p-th order Fibonacci tree of size N can be reduced to a single node in O(log N) steps on a p-th order Fibonacci cube with N nodes (processors). 2. Assume that O(log N) data items are on each of the N processors. We show that the prefix computation can be done in O(log N) steps on the p-th order Fibonacci cube.

R70-3 On Stochastic Languages

A stochastic language is a set of words accepted by a probabilistic automaton with some cutpoint. The structure of the family of stochastic languages may not parallel the structure of the family of regular languages. Some stochastic languages are context-free languages which are nonregular. The basic questions of when the complement of a stochastic language is a stochastic language, or when the intersection or union of two stochastic languages is stochastic, have not been solved but have been illuminated by Turakainens work. Turakainen proves that like the context-free languages, the intersection of a stochastic language and a regular language is a stochastic language. He shows as well that the union of a stochastic language and regular language is also a stochastic language. All right derivatives of a stochastic language are stochastic languages. In the same volume as Turakainens paper, Nasu and Honda have proven what amounts to the fact that the reversal of a stochastic language is a stochastic language. They also show other basic properties.