|Publication Type:||Conference Paper|
|Year of Publication:||2018|
|Auteurs:||Liu, S, Liu, M, Wang, M, Ma, T, Qing, X|
|Conference Name:||2018 10th International Conference on Wireless Communications and Signal Processing (WCSP)2018 10th International Conference on Wireless Communications and Signal Processing (WCSP)|
|Conference Location:||Hangzhou, China|
Vocal communication is a primary communication method of cetaceans. Learning the language of them is of great significance for the protection of cetaceans, and it also provides support for acoustic study of cetaceans’ biological behavior. In this study, a classification method based on deep learning is proposed for the classification of cetacean whistles. Firstly, the method performs short-time Fourier transform on whistles to obtain the time-frequency distribution, and then uses the deep learning model, convolutional neural network, to classify the signal based on time-frequency characteristics. The simulation results show that this method has a relatively high classification accuracy when choosing the appropriate parameters, indicating the method is effective.
Classification of Cetacean Whistles Based on Convolutional Neural Network