Benchmarking Reliability of Deep Learning Models for Pathological Gait Classification
This work addresses the problem of early detection of neurodegenerative disorders like Parkinson's for medical applications, but it is incremental as it builds on existing methods to improve reliability.
The paper analyzed existing deep learning approaches for pathological gait classification, identifying gaps that hinder practical translation, and proposed a new baseline model called Asynchronous Multi-Stream Graph Convolutional Network (AMS-GCN) that reliably differentiates multiple categories of pathological gaits across datasets.
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.