Computer-Assisted Design of Image Classification Algorithms: Dynamic and Static Fitness evaluations in a Scaffolded Genetic Programming Environment
The MIT Press eBooks · 1996
This paper discusses several issues in applying genetic programming to image classification problems in geoscience and remote sensing. In particular, this paper examines the role in using dynamic and static fitness evaluation functions. This paper also examines a few of the aspects in human-computer interactions that facilitate computer-assisted learning and problem solving (i.e., scaffolding) for our system. We describe a possible means for visualizing and summarizing a solution space without having to resort to an exhaustive search of individuals.