Systems and methods for continuous biometric authentication
摘要
說明書
The confidence level can include results from only the current sample or the confidence level can be inclusive of the entire series of interactions in the session up to the current interaction in the session (e.g., averaged). When computing the confidence level using the latter method, if the combined confidence from the first several interactions in the series is low (e.g., 2 on a scale of 1-10), then the verification must score a high confidence level in the next several interactions for the user to receive access to view any information or to participate in any activities. In some embodiments, the interactions may be weighted such that some interactions are more influential on the overall confidence level. For example, an analysis of the user speaking the user's member number may be given more weight than an analysis of the user stating the general purpose of the call since the user has likely stated the user's member number during previous interactions, resulting in a more accurate voice analysis. The various weightings may vary by industry, business, organization, user, or business unit.
