Colônia de Formigas e Busca Local para Definição Automática de Limiar na Segmentação de Imagens

• Diogo A. B. Pereira UTFPR
• Ana Cristina B. K. Vendramin UTFPR
• Alexandre de A. P. Pohl UTFPR

Resumo

This paper presents an image segmentation algorithm based on the ant colony optimization metaheuristic and on a local search. Clustering is used to group image pixels and determines the cluster centers. The local search is applied at the end of each ant solution to improve its quality. Once obtained the best centers the thresholding technique is applied to segment the image. Tests on images corroborate the performance of the metaheuristic with the local search proposed in terms of the segmentation quality.