In most conventional pattern recognition methods, the first step is extracting features from objects. These features are always expressed in the form of feature vectors. Then, the distribution of feature vectors is estimated for each category. Finally, an unknown input pattern is assigned to the category with the maximum probability. In this work, we present a system that is able to recognize objects according to the likeness of feature vectors. A database, which consists of images that were identified from a vector of related features, is used by the system to discover these objects. If the resulting image features are compared with the ones included in the database, we obtain the object that has the highest similarity with the one we proposed. The probability of success using our feature vector has been very high.
Ruiz, D., Pujol, F., García Chamizo, J., Pujol, M. and Rizo, R. (2004), "Estimation of feature vectors in object recognition processing", Kybernetes, Vol. 33 No. 1, pp. 133-140. https://doi.org/10.1108/03684920410514562Download as .RIS
Emerald Group Publishing Limited
Copyright © 2004, Emerald Group Publishing Limited