Selective security masking within recorded speech
First Claim
Patent Images
1. A method comprising:
- monitoring a call comprising speech signals between a person and an agent of a business entity using a time base;
selecting an area of a user interface that accepts special information from the person during the call;
designating at least two offset distances around the area;
identifying an activity comprising at least one of a key stroke and mouse position in at least one of the offset distances and identifying a time associated with the activity using a further time base;
measuring a time offset between the time base and the further time base;
determining a correlation coefficient for a correlation between the activity and possible entry of the special information in the area based on a location of the activity in at least one of the offset distances in comparison with a location of the area;
applying a threshold to the correlation coefficient;
determining the entry of the special information and identifying a segment of the speech signals that corresponds to the special information within the call when the correlation coefficient satisfies the threshold using the time and the measured time offset; and
rendering at least a portion of the person'"'"'s special information unintelligible.
20 Assignments
0 Petitions
Accused Products
Abstract
A marker is derived from an interaction between a person and an agent of a business and the agent'"'"'s user interface. A part of a speech signal that corresponds to a portion of the person'"'"'s special information is located with the marker. The speech signal results from the interaction between the person and the agent. The part of the speech signal that corresponds to the portion of the person'"'"'s special information is rendered unintelligible.
-
Citations
15 Claims
-
1. A method comprising:
-
monitoring a call comprising speech signals between a person and an agent of a business entity using a time base; selecting an area of a user interface that accepts special information from the person during the call; designating at least two offset distances around the area; identifying an activity comprising at least one of a key stroke and mouse position in at least one of the offset distances and identifying a time associated with the activity using a further time base; measuring a time offset between the time base and the further time base; determining a correlation coefficient for a correlation between the activity and possible entry of the special information in the area based on a location of the activity in at least one of the offset distances in comparison with a location of the area; applying a threshold to the correlation coefficient; determining the entry of the special information and identifying a segment of the speech signals that corresponds to the special information within the call when the correlation coefficient satisfies the threshold using the time and the measured time offset; and rendering at least a portion of the person'"'"'s special information unintelligible. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
-
Specification