Abstract: Technologies are provided herein for adaptively learning and matching modalities of face images. A face image is obtained and a biometric signature is generated of the face image. The biometric signature is compared with a stored biometric signature obtained from data storage. Differences are determined between the biometric signature and the stored biometric signature and compared with a proximity threshold. As a result of the comparison with the proximity threshold, the identity of the person in the face image may be determined or the face image may be stored in data storage in association with the identity of the person.

Abstract: Systems and methods for processing calls to determine if the call is potentially fraudulent or unwanted. The system extracts a speech signal from an audio signal associated with a call. The system identifies audio characteristics based on analysis of the audio signal. The system generates textual transcript of the audio signal based on automatic speech recognition of the speech signal, which is used to assign text categories for the call based on an automated multi-label textual classification of the textual transcript. The system assigns audio categories for the call based on automated multi-label acoustic classification of the speech signal. The system generates an output label for the call based on a combined analysis of the text categories, the audio categories, and the audio characteristics. The language spoken during the call may be detected and used to generate the textual transcript and to assign the text categories and the audio categories.

Abstract: Techniques are described herein for processing intra- and inter-messaging platform communications, including by receiving and analyzing messages originating from one sender for distribution to a recipient, where the sender and recipient may be on a same or separate messaging platform. Clusters of such messages with similar contents or other similar characteristics are identified and categorized, such as in accordance with configuration information regarding one or both of the originating and destination messaging platforms. Based on a determination of one or more categories associated with such an identified message cluster, as well as an analysis of metadata associated with the profile of the sender of the messages, various actions may be taken with respect to such message clusters or with parties associated with such message clusters, including actions based at least in part on the configuration information.

Abstract: A system includes a camera and a computing device. The computing device receives images captured by the camera and identifies a face in the image. A quality of the face and context surrounding the face are determined as to whether they are sufficient for determining liveness of the face. If so, a narrow-face-texture probability and a broad-face-context probability that the face is live are determined. The probabilities are compared to a live threshold and a fake threshold. In response to both probabilities exceeding the live threshold, an indication that the face is live is output. In response to either the narrow-face-texture probability or the broad-face-context probability, or both, being below the fake threshold, an indication that the face is fake is output. In response to both probabilities being above the fake threshold and below the live threshold, an indication that the image includes insufficient information is output.

Abstract: Technologies are provided herein for determining if a face image is of sufficient quality for facial recognition processing. An image that contains a face image is received and analyzed to identify facial features of the face image. Distances between the facial features are determined and dimensions of regions for the facial features are calculated based on those distances. A plurality of regions in the face image are defined based on the calculated dimensions. A combined pixel contrast quality of the face image is then calculated based on a combination of separate pixel contrast qualities for the plurality of regions. The face image is determined to be of sufficient quality for facial recognition processing in response to the combined pixel contrast quality satisfying one or more criteria.

Abstract: Technologies are provided herein for selecting one or more subimages to be processed from a plurality of images received. A subset of images is identified, from the plurality of images received, that include a face image. The face image is assessed to generate assessment information regarding one or more of a size of the face image, a pose of the face in the face image, sharpness of the face image, and contrast quality of the face image. Each aspect of the assessment information is compared with corresponding criteria and, as a result of determining that the assessment information satisfies the criteria, facial recognition processing is performed on the face image.

Abstract: Embodiments are directed towards video enhancement in accordance with time constraint. An example method includes determining a time constraint for transforming low dynamic range (LDR) video content to high dynamic range (HDR) video content, processing the LDR video content to generate instructions for transforming to HDR video content in accordance with the time constraint, rendering the HDR video content based on executing the generated instructions; and producing metadata including the generated instructions for sharing.

Abstract: Systems and methods are provided for determining a candidate identity of a person in video data based on a biometric motion signature corresponding to motion of a face of the person in the video data. Motion information is obtained corresponding to motion of the face in the video data. The motion information is provided as input to a neural network and the neural network generates the biometric motion signature in response. The biometric motion signature is compared with a plurality of biometric signatures stored in data storage. The candidate identity of the person is determined as a result of a correspondence between the biometric motion signature and a stored biometric signature of the plurality of biometric signatures.

Abstract: Technologies are provided for assessing liveness of a subject presented to a set of one or more cameras. A first image and a second image depicting a face are captured during a first time period, and a third image and a fourth image depicting the face are captured during a second time period. A first variance in facial appearance is detected between the first depiction of the face and the third depiction of the face, and a second variance in facial appearance is detected between the second depiction of the face and the fourth depiction of the face. A liveliness of a person is verified based on a determination that the first variance in facial appearance corresponds in time to the second variance in facial appearance.