Detecting panic attacks from kinect recordings

• المؤلف الرئيسي: Tan, Yong Qiang
• مؤلفون آخرون: Justin Dauwels
• التنسيق: Final Year Project
• اللغة: English
• منشور في: 2015
• الموضوعات: DRNTU::Engineering::Electrical and electronic engineering
• الوصول للمادة أونلاين: http://hdl.handle.net/10356/63794
• المؤسسة: Nanyang Technological University

Panic attack is a form of anxiety and it can happen to anybody at any time. Hence, it is important to detect it at the fastest possible time to prevent unwanted consequences. In this project, Kinect were used to record 10 subjects acting out the 4 scenarios: namely normal, heart, panic and seizure. An algorithm based on C Sharp was then implemented for extraction of joint values into text files. They were then classified based on a training and test set. This was done on Weka. Anomalies were later explained based on the scatterplots using MATlab.