Feature Selection by a Genetic Algorithm. Application to Seed Discrimination by Artificial Vision

Genetic algorithms (GAs) are efficient search methods based on the paradigm of natural selection and population genetics. A simple GA was applied for selecting the optimal feature subset among an initial feature set of larger size. The performances were tested on a practical pattern recognition problem, which consisted on the discrimination between four seed species (two cultivated and two adventitious seed species) by artiÐcial vision. A set of 73 features, describing size, shape and texture, were extracted from colour images in order to characterise each seed. The goal of the GA was to select the best subset of features which gave the highest classiÐcation rates when using the nearest neighbour as a classiÐcation method. The selected features were represented by binary chromosomes which had 73 elements. The number of selected features was directly related to the probability of initialisation of the population at the Ðrst generation of the GA. When this probability was Ðxed to 0É1, the GA selected about Ðve features. The classiÐcation performances increased with the number of generations. For example, 6É25% of the seeds were misclassiÐed by using Ðve features at generation 140, whereas another subset of the same size led to 3% misclassiÐcation at generation 400. The present work shows the great potential of GAs for feature selection (dimensionality reduction) problems. 1998 SCI. ( J Sci Food Agric 76, 77È86 (1998)