PROBABILISTIC PALM REJECTION USING SPATIOTEMPORAL TOUCH FEATURES AND ITERATIVE CLASSIFICATION

• US 20180348956A1
• Filed: 07/23/2018
• Published: 12/06/2018
• Est. Priority Date: 04/14/2014
• Status: Active Grant

• Current Assignee: Carnegie Mellon University
• Original Assignee: Carnegie Mellon University
• Inventors: Schwarz, Julia, Harrison, Chris
• Granted Patent: US 10,642,419 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 3/04186 Touch location disambiguation

Abstract

The disclosed subject matter is a palm rejection technique utilizing temporal features, iterative classification, and probabilistic voting. Touch events are classified based on features periodically extracted from time windows of increasing size, always centered at the birth of the event. The classification process uses a series of decision trees acting on said features.

Claims

1. A method, comprising:

in response to receiving a notification of a touch event associated with a touch-sensitive screen of a device, initially classifying the touch event based on a set of features that are used to characterize the touch event and a relationship of the touch event to at least one other touch event associated with the touch-sensitive screen, to facilitate distinguishing between at least a first type of touch event and a second type of touch event on the touch-sensitive screen; and

periodically re-classifying the touch event over a time window centered at a time the touch event was received, wherein, for a re-classification of the re-classifying of the touch event, the time window is expanded forward and backward in time from the time the touch event was received, while maintaining the time window centered at the time, wherein, to facilitate the re-classifying of the touch event, the re-classification is determined based on the set of features, comprising features of the touch event, previous features obtained from a previous touch event associated with the touch-sensitive screen that occurred prior to the time, and subsequent features obtained from a subsequent touch event associated with the touch-sensitive screen that occurs subsequent to the time, wherein the previous touch event and the subsequent touch event occur within the time window, and wherein a final classification of the touch event is determined based on the initial classification and the re-classification.

2. The method of claim 1, wherein the first type of touch event is a palm touch on the touch-sensitive screen, and wherein the second type of touch event is a pointer touch or a finger touch on the touch-sensitive screen.

3. The method of claim 1, further comprising:

in connection with the re-classifying of the touch event, expanding the time window symmetrically an equal amount of time forward and backward from the time the touch event was received, while maintaining the time window centered at the time.

4. The method of claim 1, wherein respective classifications of the touch event, comprising the initial classification and the re-classification, contribute respective votes to respectively characterize the touch event as at least the first type of touch event or the second type of touch event, and wherein the method further comprises:

determining the final classification based on a type of classification of the touch event that has a majority of votes of a cumulation of the respective votes, wherein the type of classification is one of the first type of touch event or the second type of touch event.

5. The method of claim 1, wherein the method further comprises:

in response to the touch event being initially classified as a pointer touch on the touch-sensitive screen or a finger touch on the touch-sensitive screen, presenting feedback to a user via the touch-sensitive screen by rendering a stroke on the touch-sensitive screen, wherein the stroke corresponds to the pointer touch or the finger touch; and

in response to a cumulation of respective votes contributed by respective classifications of the touch event indicating that the touch event is to be classified as a palm touch, removing the stroke from the touch-sensitive screen, wherein the respective classifications comprise the initial classification and the re-classification.

6. The method of claim 1, wherein the set of features comprise statistics determined over a sequence of touch events, comprising the previous touch event, the touch event, and the subsequent touch event.

7. The method of claim 6, wherein respective touch events of the sequence of touch events comprise respective radiuses of respective touch areas of the respective touch events, and wherein the method further comprises:

determining the statistics based on the respective radiuses of the respective touch events.

8. The method of claim 7, wherein the determining the statistics further comprises determining the statistics based on at least one of a mean value of the respective radiuses of the respective touch events, a standard deviation value of the respective radiuses of the respective touch events, a minimum value of the respective radiuses of the respective touch events, or a maximum value of the respective radiuses of the respective touch events.

9. The method of claim 6, further comprising:

determining a statistic of the statistics based on a total number of touch events in the sequence of touch events.

10. The method of claim 6, wherein respective touch events of the sequence of touch events comprise respective centroids of respective touch areas of the respective touch events, and wherein the method further comprises:

determining the statistics based on the respective centroids of the respective touch events.

11. The method of claim 10, wherein the respective centroids comprise a centroid of the touch event and other centroids of other touch events of the respective touch events, wherein respective distances are determined between the centroid and the other centroids, wherein the determining the statistics further comprises determining the statistics based on at least one of a mean distance of the respective distances between centroid and the other centroids, a standard deviation value of the respective distances between centroid and the other centroids, a minimum distance of the respective distances between centroid and the other centroids, or a maximum distance of the respective distances between centroid and the other centroids.

12. The method of claim 1, wherein at least one of the initial classifying or the re-classifying is performed using a set of decision trees trained using the set of features for varying window sizes of the time window.

13. The method of claim 1, further comprising:

recording touch events, comprising the touch event, the previous touch event, and the subsequent touch event, indicated by the touch-sensitive screen.

14. The method of claim 1, wherein the initial classification of the touch event occurs at a first classification time after the notification of the touch event has been received.

15. The method of claim 1, further comprising:

performing the re-classification at a first re-classification time after the notification of the touch event has been received; and

performing a second re-classification at a second re-classification time after the re-classification.

16. The method of claim 1, further comprising:

determining the final classification at a defined time or in response to determining that the touch event is terminated.

17. The method of claim 1, further comprising:

determining that a second touch event associated with the touch-sensitive screen terminates prior to a first time that a first classification of the second touch event is to be performed; and

in response to determining that the second touch event is terminated prior to the first time, determining a second final classification of the second touch event based on data relating to the second touch event that is available at the termination of the second touch event.

18. A system, comprising:

means for initially classifying a touch event associated with a touch screen of a device based on a set of features that are utilized to characterize the touch event and a relationship of the touch event to at least one other touch event associated with the touch screen, in response to receiving a notification of the touch event, to facilitate distinguishing between at least a first type of touch event and a second type of touch event on the touch screen;

means for periodically re-classifying the touch event over a time window centered at a time the touch event was received, wherein, for a re-classification of the re-classifying of the touch event performed by the means for periodically re-classifying, the time window is expanded forward and backward in time from the time the touch event was received, while maintaining the time window centered at the time, wherein, to facilitate the re-classifying of the touch event, the means for periodically re-classifying the touch event determines the re-classification based on the set of features, comprising features of the touch event, previous features obtained from a previous touch event associated with the touch screen that occurred prior to the time, and subsequent features obtained from a subsequent touch event associated with the touch screen that occurs subsequent to the time, wherein the previous touch event and the subsequent touch event occur within the time window; and

means for determining a final classification of the touch event based on the initial classification and the re-classification.

19. The system of claim 18, wherein the first type of touch event is a palm touch on the touch screen, and wherein the second type of touch event is a pointer touch or a finger touch on the touch screen.

20. The system of claim 18, wherein, in connection with the re-classifying of the touch event, the time window is expanded symmetrically an equal amount of time forward and backward from the time the touch event was received, while maintaining the time window centered at the time.