A Query Driven Computer Vision System: A Paradigm for Hierarchical Control Strategies during the Recognition Process of Three Dimensional Visually Perceived Objects.
Abstract
Two necessary components of any image understanding system are on object recognizer and a symbolic scene representation. The LandScan system currently being designed is a query driven scene analyzer in which the user's natural language queries will focus the analysis to pertinent regions of the scene. This is different than-many image understanding systems which present a symbolic description of the entire scene regardless of what portions of that picture are actually of interest. In order to facilitate such a focussing strategy, the high level analysis which includes reasoning and recognition must proceed using a top-down flow of control, and the representation must reflect the current sector of interest. In order to facilitate such a focussing strategy, the high level analysis which includes reasoning and recognition must proceed using a top-down flow of control, and the representation must reflect the current sector of interest. This paper proposes the design for a goal-oriented object recognizer and a dynamic scene representation for LandScan a system to analyze aerial photographs of urban scenes. The recognizer is an ATN in which the grammar describes sequence of primitives which define objects. The Scene Model is dynamically built as the objects specified by the queries are recognized. Thus the control of the scene modelling is top-down, reflecting the user's interest in the scene. The Scene Model represents both the objects in the image and primitive spatial relations between these objects. Keywords: Computer vision, Computer architecture. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1983
- Accession Number
- ADA185687
Entities
People
- Bajesky
Organizations
- University of Pennsylvania