Catálogo de publicaciones - libros

This paper presents a combined approach for online handwriting symbols recognition. The basic idea of this approach is to employ a set of left-right HMMs to generate a new feature vector as input, and then use SNN as a classifier to finally identify unknown symbols. The new feature vector consists of global features and several pairs of maximum probabilities with their associated different model labels for an observation pattern. A recogniser based on this method inherits the practical and dynamical modeling abilities from HMM, and robust discriminating ability from SNN for classification tasks. This hybrid technique also reduces the dimensions of feature vectors significantly, complexity, and solves size problem when using only SNN. The experimental results show that this approach outperforms several classifiers reported in recent research, and can achieve recognition rates of 97.41%, 91.81% and 91.63% for digits and upper/lower case characters respectively on the UNIPEN database benchmarks.

A common problem that one could encounter in motion estimation of indoor, or yet more, of daytime outdoor scenes is that of the detection of shadows attached to their respective moving objects. The detection of a shadow as a legitimate moving region may mislead an algorithm for the subsequent phases of analysis and tracking, which is why moving objects should be separated from their shadow. This paper presents work we have done to detect moving shadows in gray level scenes in real time for visual surveillance purposes. In this work we do not rely on any a priori information regarding with color, shape or motion speed to detect shadows. Rather, we exploit some statistical properties of the shadow borders after they have been enhanced through a simple edge gradient based operation. We developed the overall algorithm using a challenging outdoor traffic scene as a “training” sequence. Secondly, we assess the effectiveness of our shadow detection method by extracting the ground truth from gray level sequences taken indoors and outdoors from different urban and highway traffic scenes.