Invariant multi-scale object categorisation and recognition

Object recognition requires that templates with canonical views are stored in memory. Such templates must somehow be normalised. In this paper we present a novel method for obtaining 2D translation, rotation and size invariance. Cortical simple, complex and end-stopped cells provide multi-scale maps of lines, edges and keypoints. These maps are combined such that objects are characterised. Dynamic routing in neighbouring neural layers allows feature maps of input objects and stored templates to converge. We illustrate the construction of group templates and the invariance method for object categorisation and recognition in the context of a cortical architecture, which can be applied in computer vision.

Integrated multi-scale architecture of the cortex with application to computer vision

Tese de dout., Engenharia Electrónica e de Computadores, Faculdade de Ciência e Tecnologia, Universidade do Algarve, 2007

Evaluation of fruit and seed diversity and characterization of carob (Ceratonia siliqua L.) cultivars in Algarve region

The genetic diversity of 15 carob ( Ceratonia siliqua L.) cultivars located in an experimental field from Algarve (Portugal) was evaluated over 7 years using 12 fruit and seed phenotypic characters, in order to characterize carob cultivars. The values of morphological traits obtained by cultiv ar were compared with those from other countries of the Mediterranean basin. Statistically significant differences were found between cultivars for al l characters which were examined, what indicates a high genetic diversity. The relationship among these characters was analyzed by principal component analysis (PCA) resulting in the separation of these cultivars classed in four groups (clusters I–IV) and in four ungrouped cultivars. A three dimension of the model was found to be significant and explained 74.5% of the total variation, in which the first component accounting for 34.6% of the total variation is dominated by fruit characters, while the second component is dominated by seed characters. Cultivars plotted on the left-lower quadrant on the space determined by principal components 1 and 2 are characterized by fruits with high seed yield more appropriated for industrial rentability. The correlation analyses established by cultivar provided a specific understanding about the way how fruit and seed characteristics correlate within each cultivar. This approach can be useful for the development of a breeding programme, aiming to increase the seed yield, seed thickness, individual and total seed weight by fruit, characteristics that are determinant to improve the industrial exploitation of carob.