An Adjustable Gaze Tracking System and Its Application for Automatic Discrimination of Interest Objects
Amalia Istiqlali Adiba, Nobuyuki Tanaka, Jun Miyake · IEEE/ASME Transactions on Mechatronics · 2015
Gaze tracking (GT) systems have attracted the attention of researchers as well as the commercial sector. Commercial systems are readily available, but they are usually fabricated at the same size. In this study, we propose a three-dimensional (3-D)-printable frame and an open-source system to fabricate a wearable GT system with low-cost configuration and reasonable performance. A 3-D printer achieves repeatable and adjustable design. The estimated price to make the GT hardware is less than €100, and the system's average accuracy is 2.58°. Our GT system has a 24-Hz sampling rate, which can analyze human interest points. Another contribution of this study is developing the automatic discrimination of interest objects using our proposed GT and machine learning. The output of this combination system is a database that consists of labeled objects. The label with the highest frequency indicates the object of the highest interest for a corresponding user. Our combination system has a 7-Hz sampling rate and a 3.8% classification error.