Identification of fish vocalizations from ocean acoustic data. (September 2016)

• Record Type: Journal Article
• Authors: Sattar, Farook; Cullis-Suzuki, Sarika; Jin, Feng
• Is Part Of: Applied acoustics. Volume 110(2016)
• Page Start: 248
• Page End: 255
• Publication Date: 2016-09
• Subjects: Ocean acoustics -- Fish vocalizations -- Identification -- High-resolution -- Descriptors; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2016.03.025
• Languages: English
• ISSNs: 0003-682X

Abstract

Highlights: This paper deals with the challenging identification problem of fish vocalization from noisy ocean acoustic data. A new high resolution descriptor has been developed based on auditory analysis. Both the proposed descriptor as well as its application in acoustic communication of Plainfin midshipman are novel. The performance has been verified using real recorded ocean acoustic data. The promising results show that our proposed method can identify Grunts/Growls with high accuracy in real noise conditions and outperforms other relevant methods. The relationship between identification accuracy and signal-to-noise ratios has been derived. The effect of water temperature on identification accuracy has also been quantified and the derived relationship equation has been cross-validated based on a context-aware prediction algorithm. Abstract: A new method for identification of fish vocalizations based on auditory analysis and support vector machine (SVM) classification is presented. In this method, high resolution features have been extracted from fish vocalization data using the amplitude modulation spectrogram (AMS) of the input signals to facilitate the identification of grunts and growls made by a highly vocal wild fish, Porichthys notatus . The comparison results made from ocean audio recordings verify the effectiveness of the proposed method in identifying various types of fish vocalizations. The relationships between signal-to-noise ratio (SNR) and ocean temperature with the accuracy of the proposed method have also been quantified. Moreover, a context-aware prediction algorithm is introduced for estimating the continuous data.