In this project, We used an open source software, Particularly Open CV library, under software development kit (SDK) of a mobile operation system platform to develop a reliable application.
Abstract
In this project, We used an open source software, Particularly Open CV library, under software development kit (SDK) of a mobile operation system platform to develop a reliable application. Aimed in first stage to perform actions on a mobile device (smartphones, tablets and others) based on recognition of head gestures, and in the seond project phase, to track in real time after head small movements. These 2 phases separeted respectivly to 2 follow projects. The main user story is putting a phone in car docking station and answer an incoming call only by shaking the head in “Yes” gesture without moving the head (and the attention) to the phone position. Or alternatively, control virtual marker on the screen using only head motion. Reliability under any mode of operation was the main requirement while considering generic development for future features like detecting eyes and therfore try to figure out where they are looking. The technology of recognition is viola jones method corporated with basic image matching techniques.
Flowchart
Motion sensors
- Using “Rotation Vector” sensor (Ready to use data from Gyro after OS SW manipulation)
- 3 approaches:
- Detect over shaking and stop for a while
- Learn averages and detect shaking according to this position
- Correct the image detection via sensors
Future developments
- Adding new features
- Fine tuning the methods for more stable result
- Using above SW solutions in different ways
- Machine learning approach – The app will Characterize the specific user & tune the algorithm constants respectively