| CPC G06F 3/013 (2013.01) [G06F 3/0485 (2013.01); G06F 3/04886 (2013.01)] | 20 Claims |
|
1. A method comprising:
detecting, via an eye tracking device of a computing device, eye gaze data comprising a plurality of eye gaze data points of an eye movement of a user, the plurality of eye gaze data points comprising X, Y, and Z eye gaze coordinates;
sampling the plurality of detected eye gaze data points at a prescribed sampling interval;
collecting the sampled plurality of eye gaze data points into window periods having a prescribed window size;
calculating, for the sampled plurality of eye gaze data points in each window period, a weighted average of the X eye gaze coordinates;
determining, by a first machine learning model, for the sampled eye gaze data points in each window period:
a) a first probability of determining a user reading activity in each window period based on the calculated weighted average of the X eye gaze coordinates of the sampled plurality of eye gaze data points in each window period;
b) a reading location of weighted Y eye gaze coordinates; and
c) optionally, a mean value of the Z eye gaze coordinates in each window period;
calculating one or more feature extraction parameters from the sampled plurality of eye gaze data points sampled in each window period; and
determining, by a second machine learning model, a second probability of determining the user reading activity in each window period, the second probability being calculated in accordance with the first probability and the weighted Y and the mean Z eye gaze coordinates output from the first machine learning model and the one or more feature extraction parameters.
|