Understanding Stick Figures.
Abstract
This paper describes a framework for a computer model of stick figure understanding in which semantic inferences are made from body postures. Stick figures of humans are shown to provide a rich environment in which to develop and test techniques and methods of visual understanding. The computer system which will implement the theory is called SKELETUN. The notion of a hierarchical organization of recognition packets running in parallel is introduced. This paradigm is applied to an example. In addition, it is shown how SKELETUN relies on interaction between model-driven and data-driven modes. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1977
- Accession Number
- ADA047817
Entities
People
- Martin Herman
Organizations
- University of Maryland