Intelligent computer vision system
A unified approach to computer vision and human perception is proposed within the context of representation theory and the spherical model. The signal distribution received by the receptive fields is expanded in terms of spherical harmonics, with the Fourier coefficients of this expansion forming a complete set of features of the signal distribution under internal representations, and the features are then processed by the neural network. Support is given for the idea of linear and nonlinear mappings between sets of neurons being cognitive behaviors such as recognition and association, by the proposal of concrete formulations of such mappings between representation and interpretation spaces for edge detection and object recognition. 35 references.
- Research Organization:
- Purdue Univ., Indianapolis, IN; NSF, Washington, DC
- OSTI ID:
- 6458686
- Journal Information:
- Int. J. Intell. Syst.; (United States), Vol. 1
- Country of Publication:
- United States
- Language:
- English
Similar Records
Intelligent robots and computer vision XII: Algorithms and techniques; Proceedings of the Meeting, Boston, MA, Sept. 7-9, 1993
Intelligent robots and computer vision VIII: Algorithms and techniques; Proceedings of the Meeting, Philadelphia, PA, Nov. 6-10, 1989. Parts 1 2