Details

Autor(en) / Beteiligte

• Nguyen, Minh N. B.
• Hasnain, Zaki
• Ming Li
• Dorff, Tanya
• Quinn, David
• Purushotham, Sanjay
• Nocera, Luciano
• Newton, Paul K.
• Kuhn, Peter
• Nieva, Jorge
• Shahabi, Cyrus

Titel

Mining Human Mobility to Quantify Performance Status

Ist Teil von

• 2017 IEEE International Conference on Data Mining Workshops (ICDMW), 2017, p.1172-1177

Ort / Verlag

IEEE

Erscheinungsjahr

2017

Quelle

IEEE/IET Electronic Library

Beschreibungen/Notizen

• Human mobility has been studied extensively in various biomedical contexts with applications in clinical rehabilitation, disease diagnosis, health risk prognosis, and general performance assessments. In this paper, we present ATOMHP (Analytical Technologies to Objectively Measure Human Performance) Kinect: a system to objectively quantify human performance using the Microsoft Kinect as a single camera sensor to capture human mobility. We explore the viability of this noninvasive performance assessment system by studying a cohort of cancer patients undergoing various therapy regimens who are assigned a performance score based on a qualitative clinical test. The ATOM-HP Kinect is a clinically usable system which consists of tools for Kinect, clinical data collection, data quality validation, and mobility feature extraction, which can be used for downstream analysis of performance. Preliminary results based on the clinical case study indicate that ATOM-HP Kinect can quantify changes in kinematic parameters, and that these features are correlated with clinically measured risk factors which could be used for early prediction of diseases, or making decision on treatment modification.