Smartphone has become a commodity device that
makes life convenient whereas its capability of user
interaction is often limited by its small touch screen.
Employing built-in accelerometer to reconstruct gesture
motions is considered as a promising solution for
gesture interaction. However, the inability to separate
unintentional movement from gesture motion, which
often misleads the gesture recognition systems, limits
the application of those acceleration-based methods.
We present AccPen, a method that using smartphone
as a pen for gesture interaction on physical surfaces,
such as walls and tables. We utilize the high frequency
portion of acceleration signals caused by friction when
using the AccPen to write on physical surfaces to detect
intentional gestures and reconstruct a representation of
the gesture trajectory for gesture recognition.We
conduct a study with 10 participants and present robust
gesture recognition with an average accuracy of
92.22% across surfaces of desktops, wall, and paper.