Inhalt des Dokuments
Micro Gestures and User Activity Analysis Using a Wearable Finger Ring
Location: Auditorium 1, TEL,
Ernst-Reuter-Platz 7, 20th floor
SPEAKER: David Laqua (Student at TU Berlin, Computer Engineering, Bachelor)
The aim of this Bachelor thesis was to study users physical activities, as well as their micro gestures based on analyzing data registered using a finger ring. The finger ring is endowed with a variety of motion sensors including accelerometer and gyroscope, as well as proximity sensors and wireless connectivity. In this study, user’s physical activities involved regular activities such as walking, typing and writing. Such the recognition can allow the machine to adapt itself based on user’s actual needs. In an off-line classification task, the classifier achieved recognition rates of up to 97% on 6 pre-defined activity classes. In addition, interpreting fingers micro gestures for interaction with a device, mainly in the form of a mouse control operation, was also studied. An accurate and fast control of the cursor was achieved using gyroscopes sensory input. In contrast, the usage of acceleration sensor data provided a slower control, but is suitable as a cursor control when treated in a joystick like application.
HOST: Hamed Ketabdar