ABSTRACT
In the contemporary realm of athletic training, integrating technology is a pivotal determinant for augmenting athlete performance and refining training outcomes. The amalgamation of multi-target visual modeling with sensor technology imparts an enriched stratum of sports training data. Subsequently, the sensor scale-space transformation accentuates the comprehensive apprehension of data across diverse scales and angles. Hence, within this manuscript, addressing the multi-target tracking intricacies during sports training and competition, we posit a framework that amalgamates the shortest path elucidated by the K shortest paths (KSP) methodology with the pose information emanating from the Alphapose network. This framework recognizes the athlete's shortest path through a convolutional neural network and KSP, followed by the amalgamation of these divergent data sources. The fusion unfolds by incorporating the athlete's pose information grounded in Alphapose, culminating in a comprehensive integration of the two data streams. Consequently, synthesizing alpha-derived athlete information precipitates the ultimate amalgamation of the two information streams. The accomplished fusion, premised on Alphapose, forms the bedrock for multi-target tracking, culminating in a feature-rich synthesis. Empirical results reveal that after integrating these information streams, the Multiple Object Tracking Accuracy (MOTA) index and Global Multiple Object Tracking Accuracy (GMOTA) index surpass those of the solitary information tracking methods, thereby furnishing a technical underpinning and a foundation for information fusion within prospective sports training analysis systems.
ABSTRACT
Overlapping phenotypic features between Early Onset Ataxia (EOA) and Developmental Coordination Disorder (DCD) can complicate the clinical distinction of these disorders. Clinical rating scales are a common way to quantify movement disorders but in children these scales also rely on the observer's assessment and interpretation. Despite the introduction of inertial measurement units for objective and more precise evaluation, special hardware is still required, restricting their widespread application. Gait video recordings of movement disorder patients are frequently captured in routine clinical settings, but there is presently no suitable quantitative analysis method for these recordings. Owing to advancements in computer vision technology, deep learning pose estimation techniques may soon be ready for convenient and low-cost clinical usage. This study presents a framework based on 2D video recording in the coronal plane and pose estimation for the quantitative assessment of gait in movement disorders. To allow the calculation of distance-based features, seven different methods to normalize 2D skeleton keypoint data derived from pose estimation using deep neural networks applied to freehand video recording of gait were evaluated. In our experiments, 15 children (five EOA, five DCD and five healthy controls) were asked to walk naturally while being videotaped by a single camera in 1280 × 720 resolution at 25 frames per second. The high likelihood of the prediction of keypoint locations (mean = 0.889, standard deviation = 0.02) demonstrates the potential for distance-based features derived from routine video recordings to assist in the clinical evaluation of movement in EOA and DCD. By comparison of mean absolute angle error and mean variance of distance, the normalization methods using the Euclidean (2D) distance of left shoulder and right hip, or the average distance from left shoulder to right hip and from right shoulder to left hip were found to better perform for deriving distance-based features and further quantitative assessment of movement disorders.
