Benchmarking Reliability of Deep Learning Models for Pathological Gait Classification

TL;DR

This paper evaluates the reliability of deep learning models for classifying pathological gait, identifying gaps in current methods, and proposing a robust baseline model for improved generalization across datasets.

Contribution

It introduces AMS-GCN, a strong baseline model that enhances the reliability and generalization of gait classification across diverse datasets.

Findings

Identified sources of errors and generalization failures in existing approaches.

Proposed AMS-GCN as a reliable classifier for multiple gait categories.

Validated the model across simulated and real Parkinson's datasets.

Abstract

Early detection of neurodegenerative disorders is an important open problem, since early diagnosis and treatment may yield a better prognosis. Researchers have recently sought to leverage advances in machine learning algorithms to detect symptoms of altered gait, possibly corresponding to the emergence of neurodegenerative etiologies. However, while several claims of positive and accurate detection have been made in the recent literature, using a variety of sensors and algorithms, solutions are far from being realized in practice. This paper analyzes existing approaches to identify gaps inhibiting translation. Using a set of experiments across three Kinect-simulated and one real Parkinson's patient datasets, we highlight possible sources of errors and generalization failures in these approaches. Based on these observations, we propose our strong baseline called Asynchronous Multi-Stream Graph Convolutional Network (AMS-GCN) that can reliably differentiate multiple categories of pathological gaits across datasets.