Automatic Bird Species Recognition
Based on Spiking Neural Network
Ricky Mohanty, Bandi Kumar Mallik and Sandeep Singh Solanki
Abstract The main focus of this paper is to analyze sound waves produced by
birds of many different kind using spectral features, which is present in an audio
recording. This data helps in to give program examination and remote checking of
the entire populace and species of birds. This can help for natural preservation of
these species. Also survival plans and actions can be studied. The sounds produced
by these birds have special acoustic features which could help us in detecting the
type of birds. In this paper, we study about a technique which will help in finding
the type of birds using the Mel Frequency Cepstral Coefficient as well as wavelet
features and classification performance is compared using two different models. The
classification of these features is done using Spiking Neural Network (SNN) and
Gaussian Mixture Model (GMM). Spiking neural network has edge over GMM in
some aspects such as less computation time and more accuracy.
Keywords Mel frequency cepstral coefficient · Discrete wavelet transform ·
Spiking neural network (SNN) and gaussian mixture model (GMM) · Bird species
recognition
1 Introduction
Sound perceived in our day-to-day life is from various organisms as found in speech
utter by human, avian sound, amphibian calls, mammalian calls, etc. These make
sounds to do meet their daily chore or any sign of danger. These sound differ from
R. Mohanty (B ) · S. S. Solanki
Department of Electronics & Communication, Birla Institute of Technology (BIT), Mesra,
Ranchi, India
e-mail: mohantyricky@gmail.com
S. S. Solanki
e-mail: sssolanki@bitmesra.ac.in
B. K. Mallik
Central Poultry Development Organisation, Eastern Region, Bhubaneswar, India
e-mail: bandim64@gmail.com
© Springer Nature Singapore Pte Ltd. 2020
V. Nath and J. K. Mandal (eds.), Nanoelectronics, Circuits
and Communication Systems, Lecture Notes in Electrical Engineering 642,
https://doi.org/10.1007/978-981-15-2854-5_30
343