Method and apparatus for determining and analyzing a location of visual interest

A method of analyzing a location of driver visual interest derived from observation of driver physiological orientation in a vehicle, the method comprising:

• obtaining data descriptive of a plurality of driver gaze direction instances, wherein each driver gaze direction instance is defined by a data set;
• processing the data using a computer-based processor;
• classifying a plurality of the driver gaze direction instances as either on-location or off-location in preparation for further analysis related to the location of driver visual interest;
• starting a data treatment process by determining off-location visual fixations based on a determination of on-location visual fixations by classifying data that is not classified as on-location visual fixation as off-location data and ignoring the off-location visual fixation;
• clustering data derived from the classified driver gaze direction instances;
• transforming each of the plurality of classified driver gaze direction instances into one of two binary values representative of whether the respective driver gaze direction instance is on-location or off-location thus forming a classification signal which determines in each sample or unit of time whether the location of driver visual interest is on-location or off-location;
• computing a glance frequency by counting the number of on- and off intervals in the classification signal;
• computing a glance duration by regarding a start and end of each continuous segment classified in the same way;
• and
• computing a compiled data stream generated by treating the classified driver gaze direction instances with a real time hybrid algorithm which runs two tandem treatments of the classified driver gaze direction instances utilizing a standard filter and, in parallel, a stringent filter.