Detecting Gait Disorders Using Machine Learning Analysis Based on Kinect and Smartwatch Data


Uygur S., Emeç M., Özcanhan M. H.

Engineering and Technology Management in Challenging Times , Ferhan Çebi, Editör, Springer Nature, Basel, ss.189-202, 2024

  • Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
  • Basım Tarihi: 2024
  • Yayınevi: Springer Nature
  • Basıldığı Şehir: Basel
  • Sayfa Sayıları: ss.189-202
  • Editörler: Ferhan Çebi, Editör
  • İstanbul Üniversitesi Adresli: Evet

Özet

Gait disorders are significant health issues arising from neurological conditions or accidents. Evaluating and monitoring these disorders objectively using traditional methods is challenging. Therefore, advanced technological tools such as the Kinect v2 Xbox One camera sensor and a smartwatch, supported by Machine Learning methods, can provide doctors with more in-depth gait analyses. This study recorded subjects’ lower and upper extremity angles while walking for a specific duration using two Kinect camera sensors. Data from a smartwatch complement the extremity angular data. The data is processed using supervised Machine Learning techniques. The models obtained by using different Machine Learning algorithms are compared. The final two models provide objective insight into gait disorders, replacing the traditional subjective observational therapies. As such, our research presents a novel tool for detecting gait disorders. Further, clinical tests are needed for a final model suitable for the real world.