Recent advancements in pattern recognition and signal processing concern theautomatic learning of data representations from labeled training samples.Typical approaches are based on deep learning and convolutional neuralnetworks, which require large amount of labeled training samples. In this work,we propose novel feature extractors that can be used to learn therepresentation of single prototype samples in an automatic configurationprocess. We employ the proposed feature extractors in applications of audio andimage processing, and show their effectiveness on benchmark data sets.
展开▼
