KNN and SVM Classification for Chainsaw sound Identification in the Forest Areas
Résumé
We present in this paper a comparative study of two classifiers, namely, SVM (support vector machine) and KNN (K-Nearest Neighbors), which we combine to MFCC (Mel-Frequency Cepstral Coefficients) in order to make possible the detection of chainsaw's sounds in a forest environment. Optimization's calculation of the relevant characteristics of the sounds recorded in the forest and the judicious choice of the key parameters of the classifiers allows us to obtain a true positive rate of 95.63% for the SVM-LOG-KERNEL and 94.02% for the KNN. The SVM-LOG-KERNEL classifier offers a better classification result and a processing time 30 times faster than KNN.
Fichier principal
Paper_70-KNN_and_SVM_Classification_for_Chainsaw_Sound.pdf (1.01 Mo)
Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...