Improving Primate Sounds Classification using Binary Presorting for Deep Learning

TL;DR

This paper presents a novel method that improves primate sound classification accuracy by using binary presorting of audio spectrograms before applying CNNs, addressing dataset quality issues.

Contribution

It introduces a generalized relabeling approach with binary pre-sorting and CNNs, enhancing classification performance on challenging wildlife audio datasets.

Findings

Significantly higher Accuracy scores achieved

Improved UAR scores over baseline models

Effective handling of weakly labeled and noisy data

Abstract

In the field of wildlife observation and conservation, approaches involving machine learning on audio recordings are becoming increasingly popular. Unfortunately, available datasets from this field of research are often not optimal learning material; Samples can be weakly labeled, of different lengths or come with a poor signal-to-noise ratio. In this work, we introduce a generalized approach that first relabels subsegments of MEL spectrogram representations, to achieve higher performances on the actual multi-class classification tasks. For both the binary pre-sorting and the classification, we make use of convolutional neural networks (CNN) and various data-augmentation techniques. We showcase the results of this approach on the challenging \textit{ComparE 2021} dataset, with the task of classifying between different primate species sounds, and report significantly higher Accuracy and UAR scores in contrast to comparatively equipped model baselines.