Abstract: A computer-implemented method of registering image data in the context of a biometric security system includes receiving an indication of image data, detecting a first point set and a second point set, based on the image data, and generating a metric of closeness between the first point set and the second point set. In addition, the computer-implemented method includes applying the metric of closeness to a transformation model to obtain a transformation parameter and transforming the first point set with respect to the second point set, based on the transformation parameter. Further, the computer-implemented method includes determining whether transformation of the first point set with respect to the second point set indicates convergence and, if so, generating an output indicative of an aligned point set.

Abstract: Minimizing the latency of on-device detection of malicious executable files, without sacrificing accuracy, by applying a machine learning model to an executable file in quantized steps. Allowing a threshold confidence level to be set to different values enables controlling the tradeoff between accuracy and latency in generating a confidence level indicative of whether the executable file includes malware.