Detecção automática de eritrócitos em peixes usando segmentação agrupada e aprendizado supervisionado

Abstract

The growth of urban areas and the population has favored the increase in pollution and consequently the contamination of river waters. This clue has aroused interest in several aspects, mainly related to the fate and possible effects that these contaminants can cause to human health. The analysis of erythrocytes in fish is an efficient mechanism to identify the presence of geneti cal terations that maybe being caused by emergent contaminants. This article presents a new proposal for automatic identification of erythrocytes in fish using SLIC segmentation approach and connected components, adjusted using supervised learning, and presenting the performance evaluation in different aspects of the image.