Augmented Reality Based Real-Time Eyeglass Selection Using Haar Cascade Classifier

The development of Augmented Reality (AR) technology is very popular, because it is very interesting and has been widely applied in everyday life. One of the biggest challenges in AR research is getting an Immersive AR environment where real objects and virtual objects seem to merge, so fast tracking capabilities are needed to look integrated. This study tries to develop a fast face tracking system, then the results are used as a marker for natural interaction media in the AR environment. Face tracking is done with the help of a camera. The Haar Cascade Classifier algorithm is chosen to be used to find facial objects and is focused on eye tracking. After the eye features can be detected, the 3D model of the glasses will appear as an AR model of the virtual glasses. This application is built using the Android operating system, which is developed using the prototype method. The results of the experiment show that the position of the face affects the detection process. The detection process can run well when the face tilt angle is 0° (standard), 15°, 35°. While when the face tilt angle is 45° and 90° the system cannot detect the object. The detection process runs well when the face object is 20-80 cm from the camera.