Automatic Recognition of Bird Species Using Human Factor Cepstral Coefficients
Identification of bird species based on their song is very important task from biodiversity point of view. In order to develop an automatic system of recognition of bird species, a system using signal processing and pattern recognition techniques has gain
- PDF / 180,764 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 94 Downloads / 233 Views
Abstract Identification of bird species based on their song is very important task from biodiversity point of view. In order to develop an automatic system of recognition of bird species, a system using signal processing and pattern recognition techniques has gained huge importance. In this paper, we compare the performance of mel frequency cepstral coefficients and human factor cepstral coefficients combined with time- and frequency-based features. Gaussian mixture models have been used for developing feature models, and maximum likelihood estimation is used for classification. Further, selective features have been used in order to increase the performance of the system. With the proposed method, a maximum accuracy of 97.72% has been achieved for a data set of ten bird species.
Keywords Bird species recognition Mel frequency cepstral coefficients Human factor cepstral coefficients Gaussian mixture modeling
1 Introduction One of the important issues that government organizations have to deal with is the conservation of biodiversity [1]. Birds constitute an important component of biodiversity. Protection of endangered species is an important task required for conservation of biodiversity. It is based on monitoring the region in order to ascertain the presence of the species. Therefore, monitoring birds is an important activity, carried out for conservation of biodiversity. Bird experts, called as ornithologists, identify the bird species from there sounds. Such tasks rely completely on the knowledge of the ornithologists. Moreover, this is time-consuming and tedious. These limitations of manual observations have A. V. Bang (&) P. P. Rege Department of Electronics and Telecommunication, College of Engineering, Pune, India e-mail: [email protected] P. P. Rege e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2018 S. C. Satapathy et al. (eds.), Smart Computing and Informatics, Smart Innovation, Systems and Technologies 77, https://doi.org/10.1007/978-981-10-5544-7_35
363
364
A. V. Bang and P. P. Rege
given rise to automatic recognition of bird species from their audio recordings. This is a typical pattern recognition problem, which requires preprocessing, feature extraction and classification. Bird songs are varied and divided into a set of hierarchical structures [2]. The basic sound of the bird song is called as ‘syllable’ that is made up of elements. A sequence of syllables is called a ‘phrase.’ A typical combination of phrases that occur repeatedly is called as bird ‘song.’
2 Literature Survey In the past, fairly good amount of work has been carried out by various researchers to automatically recognize the bird species. Anderson et al. [3] and Kogan and Margoliash [4] are the pioneers of this work. A number of researchers have, since then, used different parametric representations of the audio signals. Mel frequency cepstral coefficients (MFCC) [5–9] are the most widely used features for bird species recognition. Some of the researchers have combined MFCC with time and frequency or descripti
Data Loading...
