Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Advanced Framework for Animal Sound Classification With Features Optimization

Qiang Yang, Xiuying Chen, Changsheng Ma, Carlos M. Duarte, Xiangliang Zhang

TL;DR

This work tackles automatic animal sound classification under challenging conditions of variable durations, diverse frequency content, and noisy environments. It introduces a three-stage pipeline that combines MFCC feature matrix rearrangement, autoencoder-based dimension reduction, and an attention-enabled Bi-LSTM classifier, validated on a marine mammal and bird benchmark. Across datasets, the proposed MDRR approach yields substantial improvements over baselines, highlighting robustness to noise and variability and offering a practical tool for biodiversity monitoring. The inclusion of a dedicated benchmark dataset and thorough parameter analyses further strengthens the framework's applicability to real-world acoustic monitoring tasks.

Abstract

The automatic classification of animal sounds presents an enduring challenge in bioacoustics, owing to the diverse statistical properties of sound signals, variations in recording equipment, and prevalent low Signal-to-Noise Ratio (SNR) conditions. Deep learning models like Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) have excelled in human speech recognition but have not been effectively tailored to the intricate nature of animal sounds, which exhibit substantial diversity even within the same domain. We propose an automated classification framework applicable to general animal sound classification. Our approach first optimizes audio features from Mel-frequency cepstral coefficients (MFCC) including feature rearrangement and feature reduction. It then uses the optimized features for the deep learning model, i.e., an attention-based Bidirectional LSTM (Bi-LSTM), to extract deep semantic features for sound classification. We also contribute an animal sound benchmark dataset encompassing oceanic animals and birds1. Extensive experimentation with real-world datasets demonstrates that our approach consistently outperforms baseline methods by over 25% in precision, recall, and accuracy, promising advancements in animal sound classification.

Advanced Framework for Animal Sound Classification With Features Optimization

TL;DR

This work tackles automatic animal sound classification under challenging conditions of variable durations, diverse frequency content, and noisy environments. It introduces a three-stage pipeline that combines MFCC feature matrix rearrangement, autoencoder-based dimension reduction, and an attention-enabled Bi-LSTM classifier, validated on a marine mammal and bird benchmark. Across datasets, the proposed MDRR approach yields substantial improvements over baselines, highlighting robustness to noise and variability and offering a practical tool for biodiversity monitoring. The inclusion of a dedicated benchmark dataset and thorough parameter analyses further strengthens the framework's applicability to real-world acoustic monitoring tasks.

Abstract

The automatic classification of animal sounds presents an enduring challenge in bioacoustics, owing to the diverse statistical properties of sound signals, variations in recording equipment, and prevalent low Signal-to-Noise Ratio (SNR) conditions. Deep learning models like Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) have excelled in human speech recognition but have not been effectively tailored to the intricate nature of animal sounds, which exhibit substantial diversity even within the same domain. We propose an automated classification framework applicable to general animal sound classification. Our approach first optimizes audio features from Mel-frequency cepstral coefficients (MFCC) including feature rearrangement and feature reduction. It then uses the optimized features for the deep learning model, i.e., an attention-based Bidirectional LSTM (Bi-LSTM), to extract deep semantic features for sound classification. We also contribute an animal sound benchmark dataset encompassing oceanic animals and birds1. Extensive experimentation with real-world datasets demonstrates that our approach consistently outperforms baseline methods by over 25% in precision, recall, and accuracy, promising advancements in animal sound classification.
Paper Structure (12 sections, 6 equations, 5 figures, 1 table)

This paper contains 12 sections, 6 equations, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methods
  3. Feature Matrix Rearrangement Module
  4. Feature Matrix Dimension Reduction Module
  5. Bi-LSTM-based Classifier with Attention Module
  6. EXPERIMENTS
  7. Experimental Settings
  8. Evaluation Results
  9. Overall Evaluation
  10. Parameter Analysis
  11. Visualization of Clustering Results
  12. Conclusion

Figures (5)

  • Figure 1: Toy example of feature matrix rearrangement.
  • Figure 2: The influence of slice lengths, max dimensions, and reduced dimension of MDRR on the marine animal dataset.
  • Figure 3: The effect of autoencoder structure of MDRR on the marine animal dataset.
  • Figure 4: The effect of input shape in MDRR on the marine animal dataset.
  • Figure 5: The learned representations from our model on the marine animal dataset.