AI

Every Step of the Way: Video-based Parkinsonian Turning Step Counting

Researchers have developed a video-based system to count the number of steps taken by people with Parkinson's disease while turning. This is challenging because people with PD often shuffle or take irregular steps when turning. The new method uses a combination of computer vision techniques, including 3D human mesh recovery and motion encoding, to estimate step counts in real-world videos. It outperforms existing methods on datasets of people with PD.
Researchers have developed a video-based system to count the number of steps taken by people with Parkinson's disease while turning. This is challenging because people with PD often shuffle or take irregular steps when turning. The new method uses a combination of computer vision techniques, including 3D human mesh recovery and motion encoding, to estimate step counts in real-world videos. It outperforms existing methods on datasets of people with PD. --- Why it matters: This matters for researchers because it could lead to more accurate assessments of motor dysfunction in Parkinson's disease patients, which is a key symptom of the condition. Source: https://arxiv.org/abs/2606.27918

This article was originally published at: https://arxiv.org/abs/2606.27918