William F. Clocksin

Perception of Surface Slant and Edge Labels from Optical Flow: A Computational Approach

I present here two results from investigations of a computational theory of how the human visual system may obtain information about the physical environment directly from optically sensed velocity fields (optical flow). Previous work has shown how optical flow arises when an observer moves through the static environment. The inverse problem is investigated here: how to recover intrinsic characteristics of the environment from optical flow patterns. The first result is a method for computing the local slant of surfaces relative to the moving observer. The second result is a method for detecting and discriminating among the five types of edges that form the boundaries of surfaces. The results have been formalised as mathematical equations, a physiological model, and a computer program. Properties and predictions resulting from these formal models are discussed with regard to experimental findings in psychophysics.

Logic programming and digital circuit analysis

Abstract We show how the unique character of logic programming can be exploited for the purpose of specifying and automatically reasoning about electrical circuits. Although propositional logic has long been used for describing the truth functions of combinational circuits, the more powerful predicate calculus on which logic programming is based has seen relatively little use in design automation. Previous researchers have introduced a number of techniques similar to logic programming, but many of the useful consequences of the methodology have not been demonstrated. We describe particular consequences of using this method for writing directly executable specifications of circuits, including the use of quantified variables, verification of hypothetical states, and sequential simulation. We have used these methods to solve problems in gate assignment, specialization of standard definitions, and determination of signal flow.

An Implementation of Model-Based Visual Feedback for Robot Arc Welding of Thin Sheet Steel

Using conventional robots for metal/inert-gas (MIG) arc welding on thin (1- to 2-mm) sheet steel pressings is restricted by the difficulty of maintaining accurate fit-up and fixturing. Dimensional variations are introduced by wear of tools and fixtures, springback, and thermal distortion during welding. These variations cause errors in the position of the welding torch relative to the seam of as much as ±3.0 mm, preventing the formation of good welds. We describe an implemented system that forms and uses models for the automatic visually guided control of a MIG welding robot. The system uses its models to aid the detection of the following dimensional variations in sheet steel assemblies: gap width, standoff error, and lateral error. Lap, T, and butt joints are dealt with. By use of the model-based visual feedback techniques reported in this paper, errors in the position of the welding torch are corrected to within ±0.5 mm, which is a suitable tolerance for producing good welds in such assemblies. Furthermore, seam widths are detected to dynamically alter welding pa rameters. Artificial intelligence principles other than models are also used, including island parsing, back constraints, and pattern-directed inference.

Structural Features of Cursive Arabic Script.

We present a technique for extracting structural features from cursive Arabic script. After preprocessing, the skeleton of the binary word image is decomposed into a number of segments in a certain order. Each segment is transformed into a feature vector. The target features are the curvature of the segment, its length relative to other segment lengths of the same word, the position of the segment relative to the centroid of the skeleton, and detailed description of curved segments. The result of this method is used to train the Hidden Markov Model to perform the recognition.

Automatic segmentation of overlapping nuclei with high background variation using robust estimation and flexible contour models

We present a segmentation method that works for overlapping and closely packed nuclei in noisy images that have high variation in background intensity. The method has been tested on fluorescence in-situ hybridisation images of interphase leucocyte nuclei. Accurate segmentation is required in support of an automatic procedure for assaying telomere content on a per area per nucleus basis. The method first finds a single seed point for each nucleus that uniquely identifies that nucleus. Seed points are located by an efficient iterative mode-finding algorithm based on robust nonparametric density estimation. Acting simultaneously on all nuclei in the image, and using the seed points as origins, flexible closed contours are dilated until each nucleus is circumscribed. Unlike previous approaches, the contour equations include a repulsive term that prevents different contours from intersecting, thereby preserving the identity of nearby or overlapping nuclei, and the contour is adaptively remeshed for greater efficiency The locations of the seed points are not critical in providing an accurate segmentation. The advantage of this method from an implementation point of view is that the computation of seed points and contours is highly efficient and robust compared with alternative approaches. The method is illustrated using data from a clinical pilot study.

A method for efficiently executing Horn clause programs using multiple processors

Previous investigations have suggested the use of multiple communicating processors for executing logic programs. However, this strategy lacks efficiency due to competition for memory and communication bandwidth, and this is a problem that has been largely neglected. In this paper we propose a realistic model for executing logic programs with low overhead on multiple processors. Our proposal does not involve shared memory or copying computation state between processors. The model organises computations over the nondeterministic proof tree so that different processors explore unique deterministic computation paths independently, in order to exploit the “OR-parallelism” present in a program. We discuss the advantages of this approach over previous ones, and suggest control strategies for making it effective in practice.

Feature representation and signal classification in fluorescence in-situ hybridization image analysis

Fast and accurate analysis of fluorescence in-situ hybridization images for signal counting will depend mainly upon two components: a classifier to discriminate between artifacts and valid signals of several fluorophores (colors), and well discriminating features to represent the signals. Our previous work (2001) has focused on the first component. To investigate the second component, we evaluate candidate feature sets by illustrating the probability density functions and scatter plots for the features. The analysis provides first insight into dependencies between features, indicates the relative importance of members of a feature set, and helps in identifying sources of potential classification errors. Class separability yielded by different feature subsets is evaluated using the accuracy of several neural network-based classification strategies, some of them hierarchical, as well as using a feature selection technique making use of a scatter criterion. Although applied to cytogenetics, the paper presents a comprehensive, unifying methodology of qualitative and quantitative evaluation of pattern feature representation essential for accurate image classification. This methodology is applicable to many other real-world pattern recognition problems.

Multi-font Arabic word recognition using spectral features

We present a new technique for recognising Arabic cursive words from scanned images of text. The approach is segmentation-free, and is applied to four different Arabic typeface, where ligatures and overlaps pose challenges to segmentation-based methods. We first transform each word into a normalised polar image, then we apply a two dimensional Fourier transform to the polar image. The resultant spectrum tolerates variations in size and rotation of displacement. Each word is represented by a template that includes a set of Fourier coefficients. The recognition is based on a normalised Euclidean distance from those templates.

Principles of the Delphi parallel inference machine

Many researchers have proposed using multiple communicating processors with a globally addressed memory for executing logic programs. However, the memory and communication contention issues characteristic of such designs have been neglected or palliated. In this paper we discuss a mew method for executing logic programs with low overhead on multiple processors. This method does not involve the sharing of memory or the copying of computation state between processors. We discuss the advantages of this approach over previous ones, and report progress on a multiprocessor implementation.

Spectral features for Arabic word recognition

We present a holistic technique for recognising words written in cursive Arabic script that does not rely on character segmentation. Each word is transformed into a normalised polar image, and a two dimensional Fourier transform is applied to the polar image. The resultant spectrum tolerates variations in size, rotation or displacement. Each word is represented by a single template, and the recognition is based on the Euclidean distance from those templates. Words are written in four different Arabic type-faces, where ligatures and overlaps pose challenges to segmentation-based methods.