Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Sound Classification of Four Insect Classes

Yinxuan Wang, Sudip Vhaduri

TL;DR

This work tackles insect sound classification among four species by leveraging MFCC features and a systematic data-augmentation pipeline to enhance generalization. It compares five classifiers—Decision Tree, Random Forest, k-NN, SVM RBF, and XGBoost—across multiple dataset configurations, including segment lengths and training/test splits that address potential leakage. Findings indicate data augmentation generally improves performance, with XGBoost delivering the strongest results under separated-clips conditions, while certain configurations reveal overfitting risks when original clips are not properly separated. The study contributes a practical analysis framework for MFCC-based insect acoustics and highlights the importance of robust data partitioning for reliable real-world deployment.

Abstract

The goal of this project is to classify four different insect sounds: cicada, beetle, termite, and cricket. One application of this project is for pest control to monitor and protect our ecosystem. Our project leverages data augmentation, including pitch shifting and speed changing, to improve model generalization. This project will test the performance of Decision Tree, Random Forest, SVM RBF, XGBoost, and k-NN models, combined with MFCC feature. A potential novelty of this project is that various data augmentation techniques are used and created 6 data along with the original sound. The dataset consists of the sound recordings of these four insects. This project aims to achieve a high classification accuracy and to reduce the over-fitting problem.

Sound Classification of Four Insect Classes

TL;DR

This work tackles insect sound classification among four species by leveraging MFCC features and a systematic data-augmentation pipeline to enhance generalization. It compares five classifiers—Decision Tree, Random Forest, k-NN, SVM RBF, and XGBoost—across multiple dataset configurations, including segment lengths and training/test splits that address potential leakage. Findings indicate data augmentation generally improves performance, with XGBoost delivering the strongest results under separated-clips conditions, while certain configurations reveal overfitting risks when original clips are not properly separated. The study contributes a practical analysis framework for MFCC-based insect acoustics and highlights the importance of robust data partitioning for reliable real-world deployment.

Abstract

The goal of this project is to classify four different insect sounds: cicada, beetle, termite, and cricket. One application of this project is for pest control to monitor and protect our ecosystem. Our project leverages data augmentation, including pitch shifting and speed changing, to improve model generalization. This project will test the performance of Decision Tree, Random Forest, SVM RBF, XGBoost, and k-NN models, combined with MFCC feature. A potential novelty of this project is that various data augmentation techniques are used and created 6 data along with the original sound. The dataset consists of the sound recordings of these four insects. This project aims to achieve a high classification accuracy and to reduce the over-fitting problem.

Paper Structure

This paper contains 37 sections, 21 figures, 13 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Bag of On-Phone ANNs to Secure IoT Objects Using Wearable and Smartphone Biometrics
  4. Insect Classification Solutions based on Insect Images
  5. Insect Classification Solutions based on Sounds
  6. Other Insect Classification Solutions
  7. Summary
  8. Methodologies
  9. Dataset
  10. ESC-50
  11. InsectSet32: Datasets for Automatic Acoustic Identification of Insects (Orthoptera and Cicadidae)
  12. Bug Bytes Sound Library: Stored Product Insect Pest Sounds
  13. Experimental Characterization and Automatic Identification of Stridulatory Sounds Inside Wood – Supplementary Information (Data)
  14. Data Segmentation
  15. Data Augmentation
  16. ...and 22 more sections

Figures (21)

  • Figure 1: Figures of Four Insect Classes
  • Figure 2: The box plot of segment durations in each class.
  • Figure 3: The bar plot of instances in each class.
  • Figure 4: t-SNE plot of instances in four classes.
  • Figure 5: UMAP plot of instances in four classes.
  • ...and 16 more figures