Despite its paramount importance for manifold use cases (e.g., in the health care industry, sports, rehabilitation and fitness assessment), sufficiently valid and reliable gait parameter measurement is still limited to high-tech gait laboratories in large clinics. Here, we demonstrate the excellent validity and test-retest repeatability of a novel gait assessment system which is built upon modern convolutional neuronal networks to extract three-dimensional skeleton joints from monocular frontal-view videos of walking humans. The validity study is based on a comparison to the GAITRite pressure-sensitive walkway system. All measured gait parameters (gait speed, cadence, step length and step time) showed excellent concurrent validity for multiple walk trials at normal and fast gait speeds. The test-retest-repeatability is on the same level as the GAITRite system. In conclusion, we are convinced that our results can pave the way for cost, space and operationally effective gait analysis in broad mainstream applications. Most sensor-based systems are costly, must be operated by extensively trained personnel (e.g., motion capture systems) or - even if not quite as costly - still possess considerable complexity (e.g. wearable sensors). In contrast, a video sufficient for the assessment method presented here can be obtained by anyone, without much training, via a smartphone camera.
Despite its paramount importance for manifold use cases, sufficiently valid and reliable gait parameter measurement is still limited to high-tech gait laboratories in large clinics. This is largely because the majority of gold-standard assessment tools are very costly and highly complex in their setup and daily operations. Here, we demonstrate the excellent validity and test-retest repeatability of a novel gait assessment system which is built upon modern convolutional neuronal networks to extract three-dimensional skeleton joints from monocular frontal-view videos of walking humans. The present validity study is achieved in comparison to a previously validated pressure-sensitive walkway system (GAITRite, GS). All measured gait parameters showed excellent level of concurrent validity. This is proven by Inter-Class-Correlations possessing values between 0.958 and 0.987 for multiple walk trials, at normal and fast gait speeds. Furthermore, the average measure of difference between the two systems is below 3.5% of corresponding gait parameter mean value across all measured parameters (0.04% - 3.25%). The percentage error values of the assessed system in relation to GS are between 5% and 13.5% of corresponding gait parameter mean values, hence being significantly below the threshold of clinical acceptability (30%). The test-retest-repeatability yields ICC values between 0.915 and 0.950, being on the same level with the GS system. In conclusion, we are convinced that our results can pave the way for cost, space and operation effective gait analysis in the broad mainstream. Most sensor-based systems are costly, have to be operated by extensively trained personnel or possess considerable complexity (e.g. wearable sensors). In contrast, a sufficient video for the assessment method presented here can be acquired by anyone, without much training, via a smartphone camera.
Despite its paramount importance for manifold use-cases, a sufficiently valid and reliable gait parameter measurement is still limited to the domain of high-tech gait laboratories in big clinics. This is mainly because the majority of gold standard assessment tools are very costly and complex in their respective setup and daily operation routines. Here, we demonstrate the excellent validity and test-retest repeatability of a novel gait assessment system which is built upon modern convolutional neuronal networks to extract three-dimensional skeleton joints from monocular frontal-view videos of walking humans. The present validity study is achieved in comparison to a previously validated pressure-sensitive walkway system (GAITRite, GS). All measured gait parameters showed excellent level of concurrent validity. This is proven by Inter-Class-Correlations possessing values between 0.92 and 0.99 for multiple walk trials, at normal and fast gait speeds. Furthermore, the average measure of difference between the two systems is below 5% of corresponding gait parameter mean value across all measured parameters (0.2% - 4.5%). The percentage error values of the assessed system in relation to GS are between 6% and 14% of corresponding gait parameter mean values, hence being significantly below the threshold of clinical acceptability (30%). The test-retest-repeatability yields ICC values between 0.87 and 0.95, being on the same level with the GS system. In conclusion, we are convinced that our results can pave the way for cost, space and operation effective gait analysis in the broad mainstream. Most sensor-based systems are costly, have to be operated by extensively trained personnel or possess considerable complexity (e.g. wearable sensors). In contrast, a sufficient video for the assessment method presented here can be acquired by anyone, without much training, via a smartphone camera.