Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Audio Classification of Low Feature Spectrograms Utilizing Convolutional Neural Networks

Noel Elias

TL;DR

This paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal.

Abstract

Modern day audio signal classification techniques lack the ability to classify low feature audio signals in the form of spectrographic temporal frequency data representations. Additionally, currently utilized techniques rely on full diverse data sets that are often not representative of real-world distributions. This paper derives several first-of-its-kind machine learning methodologies to analyze these low feature audio spectrograms given data distributions that may have normalized, skewed, or even limited training sets. In particular, this paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal. Utilizing these novel convolutional architectures as well as the proposed classification methods, these experiments demonstrate state-of-the-art classification accuracy and improved efficiency than traditional audio classification methods.

Audio Classification of Low Feature Spectrograms Utilizing Convolutional Neural Networks

TL;DR

This paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal.

Abstract

Modern day audio signal classification techniques lack the ability to classify low feature audio signals in the form of spectrographic temporal frequency data representations. Additionally, currently utilized techniques rely on full diverse data sets that are often not representative of real-world distributions. This paper derives several first-of-its-kind machine learning methodologies to analyze these low feature audio spectrograms given data distributions that may have normalized, skewed, or even limited training sets. In particular, this paper proposes several novel customized convolutional architectures to extract identifying features using binary, one-class, and siamese approaches to identify the spectrographic signature of a given audio signal. Utilizing these novel convolutional architectures as well as the proposed classification methods, these experiments demonstrate state-of-the-art classification accuracy and improved efficiency than traditional audio classification methods.

Paper Structure

This paper contains 15 sections, 1 equation, 7 figures.

Table of Contents

  1. Introduction
  2. Background
  3. Problem
  4. Related Work
  5. Approach
  6. Overall Design
  7. Binary Classification
  8. Single Class Classification
  9. Siamese Network Classification
  10. Experiments & Results
  11. Binary Classification
  12. One-Class Classification
  13. Siamese Network Classification
  14. Key Takeaways & Implementation Decisions
  15. Conclusion

Figures (7)

  • Figure 1: This illustrates algorithm to convert an audio signal to a spectrographic representation by applying the FFT on windowed portions and summing these values together. sig2spec
  • Figure 2: This illustrates the preprocessed spectrograms for sample audio signals.
  • Figure 3: This illustrates the proposed Spec-CNN architecture
  • Figure 4: This illustrates the proposed OC-SpecCNN architecture
  • Figure 5: Overview of Siamese network architecture. khandelwal_2021
  • ...and 2 more figures