This paper is concerned with the representation of objects by computer. It presents a compact relational structure representation derived by generalizing the properties of objects. Objects are described by means of surface primitives (basic descriptive elements) and relations between the primitives. The essence of the representation is the use of a single primitive to describe all identical parts. The representation involves abstracting properties of individual scene elements. Similarly, relations between parts of objects that hold for particular instances of the objects are generalized to ensure that they will be valid for all instances. Common primitives and relations are shared both within and across models. This leads to a compact representation that has many desirable properties. A computer implementation of a version of the representation has been developed. It is used to illustrate the descriptive power of the representation. A recognition system has been developed that takes advantage of the representation in its operation. It is described elsewhere (Shneier (1977), Shneier (1978)).
[1]
Stephen Alan Underwood.
Visual learning and recognition by computer
,
1972
.
[2]
Harry G. Barrow,et al.
A Versatile System for Computer-Controlled Assembly
,
1975,
Artif. Intell..
[3]
Gunnar Rutger Grape.
Model based (intermediate-level) computer-vision
,
1973
.
[4]
Ann Patricia Fothergill,et al.
Forming Models Of Plane-And-Cylinder Faceled Bodies From Light Stripes
,
1975,
IJCAI.
[5]
Michael Shneier,et al.
Object representation and recognition in machine vision
,
1979
.
[6]
Michael Shneier,et al.
Recognition Using Semantic Constraints
,
1977,
IJCAI.