Targeting saccade landing prediction using visual history
AssigneeTobii AB
InventorDaan Nijs
Life of this patent
Estimating the term…
Abstract
According to the invention, techniques for refining a ballistic prediction are described. In an example, an eye tracking system may record images over time of content presented on a display. Saccade data may be received and used as a trigger to retrieve particular ones of the recoded images. The eye tracking system may compare the images to identify a change in the content. The location of this change may correspond to a sub-area of the display. The output of the ballistic prediction may include a landing point that represents an anticipated gaze point. This landing point may be adjusted such that a gaze point is now predicted to fall within the sub-area when the change is significant.
Claims19
A method implemented by an eye tracking system, the method comprising:
- recording, in storage, a first image and a second image of content presented at a display;
- receiving saccade data indicating a user saccade;
- accessing, from the storage and based on receiving the saccade data, the first image and the second image;
- comparing the first image and the second image accessed from the storage to determine a change in the content;
- predicting, based on the saccade data and the change in the content, a gaze point on the display,
- receiving a prediction for the gaze point;
- determining, based on the comparing, a location of the change, a second change in the content, and a second location of the second change;
- determining a first distance between the location of the change and the user saccade and a second distance between the second location of the second change and the user saccade;
- selecting the change based on a comparison of the first distance and the second distance;
- and
- updating, based on the selecting, the prediction based on the location of the change such that the gaze point is predicted to be within a sub-area of the display corresponding to the location.
One or more non-transitory computer-readable storage media comprising instructions, that upon execution on a system, cause the system to perform steps comprising:
- recording, in storage, a first image and a second image of content presented at a display;
- receiving saccade data indicating a user saccade;
- accessing, from the storage and based on receiving the saccade data, the first image and the second image;
- comparing the first image and the second image accessed from the storage to determine a change in the content;
- predicting, based on the saccade data and the change in the content, a gaze point on the display;
- receiving a prediction for the gaze point;
- determining, based on the comparing, a location of the change, a second change in the content, and a second location of the second change;
- determining a first distance between the location of the change and the user saccade and a second distance between the second location of the second change and the user saccade;
- selecting the change based on a comparison of the first distance and the second distance;
- updating, based on the selecting, the prediction based on the location of the change such that the gaze point is predicted to be within a sub-area of the display corresponding to the location.
A system comprising:
- one or more processors;
- and
- one or more non-transitory computer-readable storage media comprising instructions, that upon execution by the one or more processors, configure the system to:
- record, in storage, a first image and a second image of content presented at a display;
- receive saccade data indicating a user saccade;
- access, from the storage and based on receiving the saccade data, the first image and the second image;
- compare the first image and the second image accessed from the storage to determine a change in the content;
- and
- predict, based on the saccade data and the change in the content, a gaze point on the display;
- receive a prediction for the gaze point;
- determine, based on the comparing, a location of the change, a second change in the content, and a second location of the second change;
- determine a first distance between the location of the change and the user saccade and a second distance between the second location of the second change and the user saccade;
- select the change based on a comparison of the first distance and the second distance;
- update, based on the selecting, the prediction based on the location of the change such that the gaze point is predicted to be within a sub-area of the display corresponding to the location.
Dependencies are inferred from the claim text.