Acoustic recognition of multiple bird species based on penalised maximum likelihood

Abstract

Automatic system for recognition of multiple bird species in audio recordings is presented. Time-frequency segmentation of the acoustic scene is obtained by employing a sinusoidal detection algorithm, which does not require any estimate of noise and is able to handle multiple simultaneous bird vocalisations. Each segment is characterised as a sequence of frequencies over time, referred to as a frequency track. Each bird species is represented by a hidden Markov model that models the temporal evolution of frequency tracks. The decision on the number and identity of bird species in a given recording is obtained based on maximising the overall likelihood of the set of detected segments, with a penalisation applied for increasing the number of bird models used. Experimental evaluations used audio field recordings containing 30 bird species. The presence of multiple bird species is simulated by joining the set of detected segments from several bird species. Results show that the proposed method can achieve recognition performance for multiple bird species not far from that obtained for single bird species, and considerably outperforms majority voting methods.

Automatic frequency feature extraction for bird species delimitation

“…We employed the sinusoidal detection algorithm introduced by Jančovič et al in [14], which was used in a number of works on analysis of bird vocalisations and bird species recognition, e.g., [11], [15], [16]. This method performs the detection in the short-time spectral domain.…”

Automatic detection of bird species from audio field recordings using HMM-based modelling of frequency tracks

“…In a case of tonal bird vocalisations, the use of a sinusoidal detection for segmentation also offers a natural way of representing the segment as a temporal sequence of the frequencies of the detected sinusoid, which we refer to as frequency track. This representation was employed in a few earlier studies [1], [6] and also in our recent works [3], [4], [7], [8], [9], [10]. Among the acoustic modelling approaches, the most commonly used are Gaussian mixture models (GMM) [1], [3], hidden Markov models (HMMs) [1], [4], [6], [11], and decision trees [12].…”

HMM-based modelling of individual syllables for bird species recognition from audio field recordings

“…The works in [1,2,3] employed the sinusoidal decomposition method proposed in [7]. We proposed in [8] a probabilistic method for the detection of sinusoids and employed this in our recent studies in bird pattern processing [4,5,6,9] and also here.…”

“…Although this provides a single feature vector, usually of a low dimensionality, it may not be able to describe well more complex types of syllables and may be susceptable to any variations in segmentation. In few other studies, including our recent works, [1,14,4,5,6,9], the segments were obtained based on sinusoidal detection and then represented as a temporal sequence of frequencies, which we here refer to as frequency track. The frequency track features, if extracted well, have a good potential, especially, in processing field recordings of bird vocalisations which usually contain various background noise and often also other birds/animals vocalising concurrently.…”