Detecting Gait Anomalies Based on the Analysis of Foot Pressures

The modern professional medical devices for gait analysis are very advanced but are not portable and available everywhere. However, the current information about our gait (“permanent” monitoring) might be helpful to improve our health condition. This paper describes the idea of gait anomalies detection by algorithms of deep learning. These algorithms analyze data collected by a prototype measuring device managed by a mobile application. Sensors collect foot pressure values as data for deep learning. The problem of collecting real data (from real patients) is discussed. The potential use of the device described is also discussed. Finally, assumptions and expectations regarding the results were compared with several related projects of deep learning and gait analysis. Our analysis of related studies shows that the use of deep neural networks (convolutional and recurrent) can have a satisfactory probability of detecting gait anomalies.