System and method for body language analysis
View Patent ↗A system and method are presented for body language analysis of a video interaction. In a contact center system, the video interaction between an agent and a customer may be monitored and used to determine automatic actions when threshold are met and/or matches are made. Training videos comprising determined metrics may be used for comparison to real-time interactions. Scoring and/or matches may be utilized to determine a threshold to trigger pre-determined actions based on comparison to the training videos.
1. A method for analyzing gestures of one or more parties to a video interaction in a contact center system, wherein the contact center system comprises at least a video stream analyzer, and performing actions based on the gestures, the method comprising the steps of:
a. receiving the video interaction in the contact center system from a first user to a second user, wherein the video interaction occurs through a video stream comprising a plurality of video frames comprising a plurality of pixels;
b. determining metrics, for the first user, for a previous time interval of the on-going video stream by the video stream analyzer, the determining the metrics comprising:
identifying pixels that moved in the video frames of the video stream; and
computing at least one of a motion energy image metric or a motion history image metric based on the identified pixels;
c. referencing stored metrics of system training videos, by the video stream analyzer, and comparing the stored metrics with the determined metrics of step (b) of the on-going video stream; and
d. determining if a match is made between the determined metrics and the stored metrics from the comparing of step (c), wherein if the match is made, pre-classifying the video interaction and performing an action, wherein if the match is not made, repeating the process from step (b);
wherein steps (b)-(d) cyclically occur continuously in real-time throughout a duration of the video interaction.
2. The method of claim 1 , wherein the video interaction comprises a video chat.
3. The method of claim 1 , wherein the system training videos comprise gestures indicating emotion and a plurality of levels of said emotion based on gestures.
4. The method of claim 1 , wherein the system training videos comprise gestures used in relation to at least one of: speech amplitude and keyword spotting.
5. The method of claim 1 , wherein the determining the metrics comprises determining the metrics on a cumulative basis with each passing time interval.
6. The method of claim 1 , wherein the pre-classifying comprises assigning a match to the video interaction and the action performed is based on the pre-classifying.
7. The method of claim 1 , wherein the action comprises at least one of: alerting a supervisor, alerting an authoritative figure, altering script used by the second user, recording the video interaction, generating an email to the first user, impacting the second user's performance evaluation, terminating the video interaction, placing the video interaction on hold, altering the second user's utilization score, and altering a number of interactions routed to the second user for a time interval.
8. The method of claim 1 , wherein the first user comprises a customer.
9. The method of claim 1 , wherein the second user comprises a customer service representative.
10. The method of claim 1 , wherein the previous time interval of the on-going video stream is a fraction of a second in length, and wherein the motion energy image metric or the motion history image metric is determined at least once during each such time interval.
11. A method for analyzing gestures of one or more parties to a video interaction in a contact center system, wherein the contact center system comprises at least a video stream analyzer, and performing actions based on the gestures, the method comprising the steps of:
a. receiving the video interaction in the contact center system from a first user to a second user, wherein the video interaction occurs through a video stream comprising a plurality of video frames comprising a plurality of pixels;
b. determining metrics, for at least one of the first user and the second user, for a previous time interval of the on-going video stream by the video stream analyzer, the determining the metrics comprising:
identifying pixels that moved in the video frames of the video stream; and;
computing at least one of a motion energy image metric or a motion history image metric based on the identified pixels;
c. referencing stored metrics of system training videos, by the video stream analyzer, and comparing the stored metrics with the determined metrics of step (b) of the on-going video stream;
d. updating a cumulative score in response to determining a match from the comparing the stored metrics with the determined metrics of step (c); and
e. determining if a threshold is met by the cumulative score, wherein if the threshold is met, performing an action, wherein if the threshold is not met, repeating the process from step (b), wherein
steps (b)-(d) cyclically occur continuously in real-time throughout the duration of the video interaction.
12. The method of claim 11 , wherein the threshold is determined by cumulative points from negative gestures.
13. The method of claim 11 , wherein the threshold is determined by cumulative points from negative gestures and positive gestures, wherein points from positive gestures negate points from negative gestures.
14. The method of claim 11 , wherein the video interaction comprises a video chat.
15. The method of claim 11 , wherein the system training videos comprise gestures indicating emotion and a plurality of levels of said emotion based on gestures.
16. The method of claim 11 , wherein the system training videos comprise gestures used in relation to at least one of: speech amplitude and keyword spotting.
17. The method of claim 11 , wherein the determining the metrics comprises determining the metrics on a cumulative basis with each passing time interval.
18. The method of claim 11 , wherein the action comprises at least one of: alerting a supervisor, alerting an authoritative figure, altering script used by the second user, recording the video interaction, generating an email to the first user, impacting the second user's performance evaluation, terminating the video interaction, placing the video interaction on hold, altering the second user's utilization score, and altering a number of interactions routed to the second user for a time interval.
19. The method of claim 11 , wherein the first user comprises a customer.
20. The method of claim 11 , wherein the second user comprises a customer service representative.
21. The method of claim 11 , wherein the previous time interval of the on-going video stream is a fraction of a second in length, and wherein the motion energy image metric or the motion history image metric is determined at least once during each such time interval.
22. A method for analyzing gestures of one or more parties to a video interaction in a contact center system, wherein the contact center system comprises at least a video stream analyzer, and performing actions based on the gestures, the method comprising the steps of:
a. receiving the video interaction in the contact center system from a first user to a second user, wherein the video interaction occurs through a video stream comprising a plurality of video frames comprising a plurality of pixels;
b. determining metrics, for at least one of the first user and the second user, for a previous time interval of the video stream by the video stream analyzer, the determining the metrics comprising:
identifying pixels that moved in the video frames of the video stream; and
computing at least one of a motion energy metric or a motion history metric based on the identified pixels;
c. referencing stored metrics of system training videos, by the video stream analyzer, and comparing the stored metrics with the determined metrics of step (b) of the on-going video stream; and
d. determining if one or more conditions are met from the comparing of step (c), wherein if the one or more conditions are met, pre-classifying the video interaction and performing an action, wherein if the one or more conditions are not met, repeating the process from step (b), wherein
steps (b)-(d) cyclically occur continuously in real-time throughout the duration of the video interaction.
23. The method of claim 22 , wherein the conditions comprise a match and a threshold.
24. The method of claim 23 , wherein the threshold is determined by cumulative points from negative gestures.
25. The method of claim 23 , wherein the threshold is determined by cumulative points from negative gestures and positive gestures, wherein points from positive gestures negate points from negative gestures.
26. The method of claim 22 , wherein the video interaction comprises a video chat.
27. The method of claim 22 , wherein the system training videos comprise gestures indicating emotion and a plurality of levels of said emotion based on gestures.
28. The method of claim 22 , wherein the system training videos comprise gestures used in relation to at least one of: speech amplitude and keyword spotting.
29. The method of claim 22 , wherein the determining the metrics comprises determining the metrics on a cumulative basis with each passing time interval.
30. The method of claim 22 , wherein the pre-classifying comprises assigning a match to the video interaction and the action performed is based on the pre-classifying.
31. The method of claim 22 , wherein the action comprises at least one of: alerting a supervisor, alerting an authoritative figure, altering script used by the second user, recording the video interaction, generating an email to the first user, impacting the second user's performance evaluation, terminating the video interaction, placing the video interaction on hold, altering the second user's utilization score, and altering a number of interactions routed to the second user for a time interval.
32. The method of claim 22 , wherein the first user comprises a customer.
33. The method of claim 22 , wherein the second user comprises a customer service representative.
34. The method of claim 22 , wherein the previous time interval of the on-going video stream is a fraction of a second in length, and wherein the motion energy image metric or the motion history image metric is determined at least once during each such time interval.