SUBSET SELECTION FROM BIASED DICTIONARIES FOR IMPACT ACOUSTIC CLASSIFICATION

In this paper we study a sparse signal representation ap-proach for the classification of impact acoustic signals ob-tained from empty and full hazelnuts. In particular, two cus-tom dictionaries are designed for each class with a vector quantization algorithm by using the training data for each. In the following step each individual dictionary or their combination is used for representing the test acoustic sig-nals with the belief that the representation will be biased. Two different subset selection (SS) techniques, matching pursuit (MP) and a bounded error subset selection algo-rithm (BESS) were investigated to approximate given signals by using the code vectors in these biased dictionaries. The approximation error and the number of code vectors se-lected from dictionaries were used as input features for th