GeHirNet: A Gender-Aware Hierarchical Model for Voice Pathology Classification

arXiv (Cornell University)

Abstract

AI-based voice analysis shows promise for disease diagnostics, but existing classifiers often fail to accurately identify specific pathologies because of gender-related acoustic variations and the scarcity of data for rare diseases. We propose a novel two-stage framework that first identifies gender-specific pathological patterns using ResNet-50 on Mel spectrograms, then performs gender-conditioned disease classification. We address class imbalance through multi-scale resampling and time warping augmentation. Evaluated on a merged dataset from four public repositories, our two-stage architecture with time warping achieves state-of-the-art performance (97.63\% accuracy, 95.25\% MCC), with a 5\% MCC improvement over single-stage baseline. This work advances voice pathology classification while reducing gender bias through hierarchical modeling of vocal characteristics.

Authors 4

1. Fan Wu

   ETH Zurich

   Affiliation as printed

   Centre for Digital Health Interventions, ETH Zurich, Zurich, Switzerland

2. Kaicheng Zhao Aachen Institute of Mechanism Theory

   RWTH Aachen University

   Affiliation as printed

   Institute of Mechanism Theory, Machine Dynamics and Robotics, RWTH Aachen University, Aachen, Germany

3. Elgar Fleisch

   ETH Zurich · University of St.Gallen

   Affiliation as printed

   Centre for Digital Health Interventions, ETH Zurich, Zurich, Switzerland

   Centre for Digital Health Interventions, University of St. Gallen, St. Gallen, Switzerland

4. Filipe Barata

   ETH Zurich

   Affiliation as printed

   Centre for Digital Health Interventions, ETH Zurich, Zurich, Switzerland