Explaining Machine Learning Models for Clinical Gait Analysis

Machine Learning (ML) is increasingly used to support decision-making in the healthcare sector. While ML approaches provide promising results with regard to their classification performance, most share a central limitation, their black-box character. This article investigates the usefulness of Explainable Artificial Intelligence (XAI) methods to increase transparency in automated clinical gait classification based on time series. For this purpose, predictions of state-of-the-art classification methods are explained with a XAI method called Layer-wise Relevance Propagation (LRP). Our main contribution is an approach that explains class-specific characteristics learned by ML models that are trained for gait classification. We investigate several gait classification tasks and employ different classification methods, i.e., Convolutional Neural Network, Support Vector Machine, and Multi-layer Perceptron. We propose to evaluate the obtained explanations with two complementary approaches: a statistical analysis of the underlying data using Statistical Parametric Mapping and a qualitative evaluation by two clinical experts. A gait dataset comprising ground reaction force measurements from 132 patients with different lower-body gait disorders and 62 healthy controls is utilized. Our experiments show that explanations obtained by LRP exhibit promising statistical properties concerning inter-class discriminativity and are also in line with clinically relevant biomechanical gait characteristics.

Gait Recognition System: A Survey

Human identification by gait has created a great deal of interest in computer vision community due to its advantage of inconspicuous recognition at a relatively far distance. This paper provides a comprehensive survey of recent developments on gait recognition approaches. The survey emphasizes on three major issues involved in a general gait recognition system, namely gait image representation, feature dimensionality reduction and gait classification. It also gives, a review of the available public gait datasets. The concluding discussions outline a number of research challenges and provide promising future directions for the field. In this article presents a survey of the work done in gait analysis for re-identification in the past decade, looking at the main approaches, datasets, and evaluation methodologies also discussed.

Automatic methods of hoof-on and -off detection in horses using wearable inertial sensors during walk and trot on asphalt, sand and grass

Detection of hoof-on and -off events are essential to gait classification in horses. Wearable sensors have been endorsed as a convenient alternative to the traditional force plate-based method. The aim of this study was to propose and validate inertial sensor-based methods of gait event detection, reviewing different sensor locations and their performance on different gaits and exercise surfaces. Eleven horses of various breeds and ages were recruited to wear inertial sensors attached to the hooves, pasterns and cannons. Gait events detected by pastern and cannon methods were compared to the reference, hoof-detected events. Walk and trot strides were recorded on asphalt, grass and sand. Pastern-based methods were found to be the most accurate and precise for detecting gait events, incurring mean errors of between 1 and 6ms, depending on the limb and gait, on asphalt. These methods incurred consistent errors when used to measure stance durations on all surfaces, with mean errors of 0.1 to 1.16% of a stride cycle. In conclusion, the methods developed and validated here will enable future studies to reliably detect equine gait events using inertial sensors, under a wide variety of field conditions.

Author Correction: Improving gait classification in horses by using inertial measurement unit (IMU) generated data and machine learning

An amendment to this paper has been published and can be accessed via a link at the top of the paper.