Generating and generalizing models of visual objects
Artificial IntelligencePublished 1 February 1987Open access
Jonathan H. Connell, Michael Brady
Citations191
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
An implemented learning system whose inputs are images of two-dimensional shapes is reported on, which can learn disjunctions, and can learn concepts using only positive examples, and discusses learnability and the pervasive importance of representational hierarchies.
Abstract
We report on initial experiments with an implemented learning system whose inputs are images of two-dimensional shapes. The system first builds semantic network descriptions of shapes based on Brady's "smoothed local symmetry" representation. It learns shape models from them using a substantially modified version of Winston's analogy program. A generalization of Gray coding enables the representation to be extended and also allows a single operation, called "ablation," to achieve the effects of many standard induction heuristics. The program can learn disjunctions, and can learn concepts using only positive examples. We discuss learnability and the pervasive importance of representational hierarchies.
Keywords
Computer ScienceBiochemistry, Genetics and Molecular Biology
Artificial IntelligenceDiagnostic reasoning based on structure and behavior
973 Citations1984Randall Davis
A system that reasons from first principles, i.e., using knowledge of structure and behavior, which is implemented and tested on several examples in the domain of troubleshooting digital electronic circuits and describes a technique it calls constraint suspension that provides a powerful tool for troubleshooting.
Artificial IntelligenceSymbolic reasoning among 3-D models and 2-D images
775 Citations1981Rodney A. Brooks
Modelling, prediction, description and interpretation proceed concurrently from coarse object subpart and class interpretations of images, to fine distinctions among object subclasses and more precise three dimensional quantification of objects.
IEEE Transactions on Pattern Analysis and Machine IntelligenceThe Curvature Primal Sketch
697 Citations1986Haruo Asada, Michael Brady
An implemented algorithm is described that computes the Curvature Primal Sketch by matching the multiscale convolutions of a shape, and its performance on a set of tool shapes is illustrated.
Communications of the ACMLearning and reasoning by analogy
525 Citations1980Patrick Henry Winston
A theory of analogy is presented and an implemented system that embodies the theory is described that is designed to answer questions about Hamlet by way of knowledge about Macbeth.
The International Journal of Robotics ResearchSmoothed Local Symmetries and Their Implementation
362 Citations1984Michael Brady, Haruo Asada
The main features of the SLS repre sentation are developed and an implemented algorithm that com putes it is described and illustrated for a set of tools.
Computer Vision Graphics and Image ProcessingDescribing surfaces
251 Citations1985Michael Brady, Jean Ponce +2 more
DSpace@MIT (Massachusetts Institute of Technology)Learning physical descriptions from functional definitions, examples, and precedents
198 Citations1983Patrick Henry Winston, Thomas O. Binford +2 more
The theory is a synthesis of two sets of ideas: ideas about learning from precedents and exercises developed at MIT and ideas about physical description developed at Stanford, and is illustrated by way of representative experiments.
Artificial IntelligenceInductive learning of structural descriptions
194 Citations1981Thomas G. Dietterich, Ryszard S. Michalski
Primary attention is given to methods of learning the simplest form of generalization, namely, the maximally specific conjunctive generalizations (MSC-generalizations) which completely characterize a single set of structural examples.
Artificial IntelligenceLearning new principles from precedents and exercises
158 Citations1982Patrick Henry Winston
A theory of how it is possible to learn by precedents and exercises is presented and an implemented system that exploits the theory is described that holds that causal relations identify the regularities that can be exploited from past experience.
Elsevier eBooksCriteria for Representations of Shape
125 Citations1983Michael Brady
A representation for the shape of surfaces is proposed, called Curvature Patches, used to extend Smoothed Local Symmetries to a novel three-dimensional representation.
FigshareA model for functional reasoning in design
121 Citations2018Peter Freeman, Allen Newell
A model of the design process is developed in two stages corresponding to the task environment of design and the activity of posing and solving design problems, and issues raised by the model's qualitative aspects, its suitability to automated design, and lines for further development are discussed.
Kluwer international series in engineering and computer scienceToward a Surface Primal Sketch
88 Citations1987Jean Ponce, Michael Brady
Cognitive ScienceInduction of Augmented Transition Networks*
83 Citations1977John R. Anderson
Critical to the performance of the program are assumptions that it makes about the relation between sentence structure and surface structure, about when word classes may be formed and when ATN networks may be merged, and about the structure of noun phrases.
DSpace@MIT (Massachusetts Institute of Technology)Learning by Augmenting Rules and Accumulating Censors.
66 Citations1982Patrick Henry Winston
A synthesis of several sets of ideas enables two extensions to an implemented system that solves problems involving precedents and exercises and that generates if-then rules as a byproduct.
National Conference on Artificial IntelligenceA variational approach to edge detection
54 Citations1983John Canny
An attempt to formulate a set of edge detection criteria that capture as directly as possible the desirable properties of the detector, including a one dimensional operator that approximates the first derivative of a Gaussian.
Artificial IntelligenceRelational production systems☆
32 Citations1977Steven A. Vere
A relational production composition theorem is demonstrated that without a sponge-like component in the antecedent of a production, composition of two arbitrary productions is in general impossible.
DSpace@MIT (Massachusetts Institute of Technology)Learning Shape Descriptions: Generating and Generalizing Models of Visual Objects
28 Citations1985Jonathan H. Connell
An implemented system for learning structural prototypes from gray-scale images and a learning algorithm that forms class models from the descriptions produced and used these models to recognize new members of the class is exhibited.
DSpace@MIT (Massachusetts Institute of Technology)Local Rotational Symmetries
27 Citations1985Margaret M. Fleck
First, grey-scale image smoothing proves to be better than boundary smoothing for creating representations at multiple scales of resolution, because it is more robust and it allows qualitative changes in representations between scales.
Three-Dimensional Vision
9 Citations2000Joseph H. Nurre