Intelligent Object Recognition for Manufacturing Applications
Girija Chetty, Narendra Deshpande · Sixth International Conference on Manufacturing Engineering: Manufacturing; a Global Perspective; Proceedings, The · 1995
Automatic recognition of an object involves several stages such as low level processing, feature extraction and finally model matching with extracted features. Several algorithms and procedures are now available for each one of the above stages and these tend to be highly complex as well as context dependent, because of which a suitable choice for a given application is rather dii3cult for a non-expert in this discipline. This paper proposes a semantically integrated system for recommending the appropriate algorithms for various stages in object recognition. The system proposed here is functionally divided into four different modules: Geometrical reasoning expert (GRE), Model selection expert (MSE), Low-level vision expert (LLVE) and Question and Answer Module (QAM). The reported work gives details of Low Level Vision Expert module based on semantic integration of image processing procedures paradigm. Given an image, the system suggests a suitable image processing procedure, allows the suggested procedure to be applied to the image and suggests an alternate choice in case of unsatisfactory results.