Classification vs. Regression in Supervised Learning for Single Channel Speaker Count Estimation

TL;DR

This paper compares classification and regression approaches in deep neural networks for estimating the number of concurrent speakers from single-channel audio, evaluating different input representations and architectures.

Contribution

It provides an empirical analysis of classification versus regression strategies for speaker count estimation using a Bi-LSTM DNN model.

Findings

Classification and regression approaches have different strengths for speaker count estimation.

Input representation significantly impacts model performance.

The study achieves accurate estimation for mixtures of up to ten speakers.

Abstract

The task of estimating the maximum number of concurrent speakers from single channel mixtures is important for various audio-based applications, such as blind source separation, speaker diarisation, audio surveillance or auditory scene classification. Building upon powerful machine learning methodology, we develop a Deep Neural Network (DNN) that estimates a speaker count. While DNNs efficiently map input representations to output targets, it remains unclear how to best handle the network output to infer integer source count estimates, as a discrete count estimate can either be tackled as a regression or a classification problem. In this paper, we investigate this important design decision and also address complementary parameter choices such as the input representation. We evaluate a state-of-the-art DNN audio model based on a Bi-directional Long Short-Term Memory network architecture for speaker count estimations. Through experimental evaluations aimed at identifying the best overall strategy for the task and show results for five seconds speech segments in mixtures of up to ten speakers.