Análise de forma e robustez : métodos existentes e algumas proposições

Abstract

Statistical analysis of shapes is the study of variations in shape, size and their covariations. Although most applications relate to biology, the statistical analysis of shapes has become a powerful tool, with diverse applications in the fields of archeology, paleontology, geography, and medicine. Procustes analysis is one of the most used methods in the statistical analysis of shapes and plays an important role in measuring, comparing and estimating the average shape of objects. However, such a technique is based on the least squares method, which is severely affected when the data have outliers. In this work we propose a set of methods for procustes analysis considering some of the existing robust estimators in the literature.