Visual Algorithms.
Abstract
Nonlinear, local and highly parallel algorithms can perform several simple but important visual computations. Specific classes of algorithms can be considered in an abstract way. The author here the class of polynomial algorithms to exemplify some of the important issues for visual processing like linear vs. nonlinear and local vs. global. Polynomial algorithms are a natural extension of Perceptrons to time dependent grey level images. Although they share most of the limitations of Perceptrons, they are powerful parallel computational devices. Several of their properties are characterized and especially their equivalence with Perceptrons for geometrical figures and the synthesis of nonlinear algorithms(mapping) via associative learning. Finally, the paper considers how algorithms of this type could be implemented in nervous hardware, in terms of synaptic interactions strategically located in a dendritic tree. The implementation of three specific algorithms is briefly outlined: direction sensitive motion detection; detection of discontinuities in the optical flow; and detection and localization of zero-crossings in the convolution of the image with the Laplacian (of a Gaussian). In the appendix, another (nonlinear) differential operator, the second directional derivative along the gradient, is briefly discussed as an alternative to the Laplacian. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1982
- Accession Number
- ADA127251
Entities
People
- Tomas Lozano-Pérez
Organizations
- Massachusetts Institute of Technology