Automatic Classification of Gait Impairments Using a Markerless 2D Video-Based System

Abstract

Systemic disorders affecting an individual can cause gait impairments. Successful acquisition and evaluation of features representing such impairments make it possible to estimate the severity of those disorders, which is important information for monitoring patients’ health evolution. However, current state-of-the-art systems perform the acquisition and evaluation of these features in specially equipped laboratories, typically limiting the periodicity of evaluations. With the objective of making health monitoring easier and more accessible, this paper presents a system that performs automatic detection and classification of gait impairments, based on the acquisition and evaluation of biomechanical gait features using a single 2D video camera. The system relies on two different types of features to perform classification: (i) feet-related features, such as step length, step length symmetry, fraction of foot flat during stance phase, normalized step count, speed; and (ii) body-related features, such as the amount of movement while walking, center of gravity shifts and torso orientation. The proposed system uses a support vector machine to decide whether the observed gait is normal or if it belongs to one of three different impaired gait groups. Results show that the proposed system outperforms existing markerless 2D video-based systems, with a classification accuracy of 98.8%.

Using transfer learning for classification of gait pathologies

“…The other two datasets, however, are captured in less constrained environments, where the segmentation of silhouettes is far form perfect, often missing parts of the walking person's silhouette. Thus, it can be concluded that the proposed finetuning scheme generalizes well across datasets, even in the presence of silhouette segmentation errors, which affect the performance of most systems based on biomechanical handcrafted features [1]. Table III also reports results for the state-of-the-art methods.…”

“…Most vision based systems that rely on a 2D video only perform a binary classification of whether the observed gait is normal or impaired. While some systems, such as [1], can identify gait pathologies, their results are very much dependent on the quality of the silhouettes used. In such situations, poor silhouette segmentation can significantly reduce the classification results.…”

“…The GEI method [4] performs well with the binary classification problem of the DAI Gait dataset, but its performance degrades significantly when the number of pathology groups increases. The biomechanical feature extraction method [1] that operates on INIT Gait Dataset with a classification accuracy of 98% cannot be applied to the DAI and DAI 2 datasets, since it relies on features such as shift in centre of gravity, torso orientation and amount of movement, which cannot be reliably computed in the presence of silhouete segmentation errors. Finally, the leg angle method [15] is very effective in performing binary classification over the DAI dataset but, as reported in [4], it cannot be used to perform pathology specific classification.…”

“…Gait is a highly cognitive task that involves a coordinated, cyclic combination of movements which results in human locomotion [1]. Analysis of gait can provide useful information about an individual's health, identity, gender or walking patterns, having a wide range of applications in fields such as sports, biometrics or medicine.…”

“…However, as presented in [4], features acquired from a gait representation typically used for biometric recognition, called the gait energy image (GEI), can also be used. Although the effectiveness of such systems is currently lower than the systems that rely on biomechanical features [1], their robustness to silhouettes segmentation errors makes them advantageous in a daily life setting, where the capture of gait sequences cannot be performed in (near-) ideal conditions. This paper presents a novel system that performs classification of gait across different pathologies, based on features learned by a deep convolutional neural network (CNN) network, VGG-19, after appropriately finetuning it to the problem addressed.…”