Using Machine Learning Method to Identify for Frog Classification

This paper presents the application of machine learning (ML) method to do frog classification. Bioacoustic features are filtered with respect of Mel-scale frequency cepstral coefficient (MFCC) algorithm. The raw data is collected from real frog sounds by using a digital signal processing technique. For classification and identification of frog, we can adopt the neural network (NN) of ML algorithm. Experimental results are carried out for bioacoustic features four frogs. The effectiveness of the proposed system on the bioacoustic features is identified by using the general NN scheme. Implementation results on CPU and GPU processors can be discussed.