An efficient local receptive field feature extractor for handwriting recognition
Radu Dogaru, Ioana Dogaru · 2016
The performance of a simple yet efficient local receptive field feature extractor is evaluated on state of the art handwritten databases showing that after the proper optimization of its parameters, very good accuracy performances can be obtained using a shallow classifier (e.g. the support vector machine), close to the ones achieved using more sophisticated techniques such as deep-learning classifiers. Particularly useful is the computing speed acceleration of more than one order of magnitude in both training and prediction modes, as a result of reducing the effective size of the feature vector to an order of tens inputs instead hundreds as in the case of applying the input samples directly. Consequently such feature extraction techniques can be conveniently embedded into smart sensing units with various applications from intelligent sensors to portable biometric authentication systems.