Abstract: A method, apparatus and system for human skeleton pose estimation includes synchronously capturing images of a human moving through an area from a plurality of different points of view, for each of the plurality of captured images, determining a bounding box that bounds the human in the captured image and identifying pixel locations of the bounding box in the image, for each of the plurality of captured images, determining 2D and single-view 3D skeletons from the pixel locations of the bounding box, determining a first, multi-view 3D skeleton using a combination of the 2D and single-view 3D skeletons, and optimizing the first, multi-view 3D skeleton to determine a final 3D skeleton pose for the human. The method, apparatus and system can further include illuminating the area with structured light during the capturing of the images of the human moving through the area.

Abstract: In general, techniques for data access control are described, in which a policy engine implements and applies a declarative policy framework that can represent and reason about complex privacy policies. By using a common data model together with a formal shareability theory, this declarative policy framework enables the specification of expressive policies in a concise way without burdening the user with technical details of the underlying formalism of a data querying application or other knowledge representation scheme. The policy engine may be deployed as the policy decision point in a data access control system that also includes a policy enforcement point. The policy engine includes user interfaces for the creation, validation, and management of privacy policies. The policy engine may interface with systems that manage data requests and replies by coordinating policy engine decisions and access to databases.

Abstract: A method, apparatus and system for visual grounding of a caption in an image include projecting at least two parsed phrases of the caption into a trained semantic embedding space, projecting extracted region proposals of the image into the trained semantic embedding space, aligning the extracted region proposals and the at least two parsed phrases, aggregating the aligned region proposals and the at least two parsed phrases to determine a caption-conditioned image representation and projecting the caption-conditioned image representation and the caption into a semantic embedding space to align the caption-conditioned image representation and the caption. The method, apparatus and system can further include a parser for parsing the caption into the at least two parsed phrases and a region proposal module for extracting the region proposals from the image.

Abstract: An example system can include a reference biometric template (RBT) reader, an authenticator, and an auxiliary system. In some examples, during an initial enrollment process, the RBT reader obtains a biometric from a user, transforms the biometric into an RBT, and provides different shares of the RBT to the authenticator and the auxiliary system. The authenticator and the auxiliary system create respective shares of helper data. In some examples, the authenticator and the auxiliary system use a non-commutative transformation function to embed a secret key in their respective shares of the helper data. The auxiliary system provides its share of the helper data to the authenticator. The authenticator combines its share of the helper data with the share provided by the auxiliary system to create a full version of the helper data. The helper data can be used in a subsequent authentication process between the RBT reader and the authenticator.

Abstract: Systems and methods for speech recognition are provided. In some aspects, the method comprises receiving, using an input, an audio signal. The method further comprises splitting the audio signal into auditory test segments. The method further comprises extracting, from each of the auditory test segments, a set of acoustic features. The method further comprises applying the set of acoustic features to a deep neural network to produce a hypothesis for the corresponding auditory test segment. The method further comprises selectively performing one or more of: indirect adaptation of the deep neural network and direct adaptation of the deep neural network.

Abstract: A variety of applications involve transmission of video data in real time. However, bandwidth limitations may result in degradation of the quality of the transmitted video. To address such limitations, a video stream may be limited to ‘regions of interest’ within each frame of the video stream. This can be accomplished by setting areas outside of the regions of interest to an a priori known color or ‘blanking’ color, and the resulting ‘blanked’ video stream can be compressed and transmitted without extensive modifications to existing transmission apparatus. The original video can be reconstituted by superimposing the received region-of-interest video data onto a background image of the environment that contains the regions-of-interest. The background image can be satellite imagery or other stored imagery that matches the perspective of the video stream. Geographical tag data can be used to superimpose the region-of-interest video data onto the background image.

Abstract: A method, apparatus and system for understanding visual content includes determining at least one region proposal for an image, attending at least one symbol of the proposed image region, attending a portion of the proposed image region using information regarding the attended symbol, extracting appearance features of the attended portion of the proposed image region, fusing the appearance features of the attended image region and features of the attended symbol, projecting the fused features into a semantic embedding space having been trained using fused attended appearance features and attended symbol features of images having known descriptive messages, computing a similarity measure between the projected, fused features and fused attended appearance features and attended symbol features embedded in the semantic embedding space having at least one associated descriptive message and predicting a descriptive message for an image associated with the projected, fused features.

Abstract: This disclosure describes techniques that include generating, based on a description of a scene, a movie or animation that represents at least one possible version of a story corresponding to the description of the scene. This disclosure also describes techniques for training a machine learning model to generate predefined data structures from textual information, visual information, and/or other information about a story, an event, a scene, or a sequence of events or scenes within a story. This disclosure also describes techniques for using GANs to generate, from input, an animation of motion (e.g., an animation or a video clip). This disclosure also describes techniques for implementing an explainable artificial intelligence system that may provide end users with information (e.g., through a user interface) that enables an understanding of at least some of the decisions made by the AI system.

Abstract: A correlator that includes a number of modules cooperating with each other. A transaction correlation module correlates network flow information for one or more network packet flows corresponding to one or more host-agent network-transaction records on whom participated in a network packet flow. The host-agent network-transaction records at least contain source information. A host input module to take in the host-agent network-transaction records from each host agent on its host computing device connecting to the correlator. A merged record creator creates a merged record for corresponding matches of one or more of the host-agent network-transaction records to one or more of the network packet flows. The merged record gives the network policy enforcement module a complete picture of both the network traffic flow information along with the source information that participated in the network packet flows in order to apply network polices against the network packet flows.

Abstract: An example antenna system includes a connection member, a first pair of antipodal Vivaldi antennas, and a second pair of antipodal Vivaldi antennas. The first pair of antipodal Vivaldi antennas are coupled to the connection member, positioned co-planar with each other along a first plane, and inverted relative to each other. The first pair of antennas provide approximately 180 degrees of phase shift (frequency independent) for a first group of signals. The second pair of antipodal Vivaldi antennas are coupled to the connection member, positioned co-planar with each other along a second plane substantially orthogonal to the first plane, and inverted relative to each other. The second pair of antennas provide approximately 180 degrees of phase shift (frequency independent) for a second group of signals. The antenna system is configured to utilize the first and second pairs of antennas to transmit or receive signals with circular polarization.

Abstract: Embodiments of the disclosed technologies include finding content of interest in an RF spectrum by automatically scanning the RF spectrum; detecting, in a range of frequencies of the RF spectrum that includes one or more undefined channels, a candidate RF segment; where the candidate RF segment includes a frequency-bound time segment of electromagnetic energy; executing a machine learning-based process to determine, for the candidate RF segment, signal characterization data indicative of one or more of: a frequency range, a modulation type, a timestamp; using the signal characterization data to determine whether audio contained in the candidate RF segment corresponds to a search criterion; in response to determining that the candidate RF segment corresponds to the search criterion, outputting, through an electronic device, data indicative of the candidate RF segment; where the data indicative of the candidate RF segment is output in a real-time time interval after the candidate RF segment is detected.

Abstract: Embodiments provide a stationary X-ray source for a multisource X-ray imaging system for tomographic imaging. The stationary X-ray source includes an array of thermionic cathodes and, in most embodiments a rotating anode. The anode rotates about a rotation axis, however the anode is stationary in the horizontal or vertical dimensions (e.g. about axes perpendicular to the rotation axis). The elimination of mechanical motion improves the image quality by elimination of mechanical vibration and source motion; simplifies system design that reduces system size and cost; increases angular coverage with no increase in scan time; and results in short scan times to, in medical some medical imaging applications, reduce patient-motion-induced blurring.

Abstract: A method, apparatus and system for enhanced welding visualization include splitting incoming light from the welding environment into at least a first optical path and a second optical path having different light levels using at least one beam splitter. Images of the split light having different light levels are captured using a respective imaging sensor. The images from the respective imaging sensors are fused to create a left-eye fused image and a right-eye fused image. The left-eye fused image is displayed on a display at a location of a left eye of a user and the right-eye fused image is displayed on a display at a location of a right eye of the user to provide a parallax-free, high dynamic range, representation of the welding environment.

Abstract: The disclosure includes an ionization chamber, a first electron multiplier, and a second electron multiplier. The ionization chamber is configured to receive gas molecules from an environment at a pressure. The first electron multiplier is configured to receive a plurality of photons from a photon source, generate a first plurality of electrons from the plurality of photons, and discharge the first plurality of electrons into the ionization chamber to generate a plurality of gas ions from at least a portion of the gas molecules. The second electron multiplier is configured to receive the plurality of gas ions from the ionization chamber and generate a second plurality of electrons from the plurality of gas ions that is proportional to a quantity of the plurality of gas ions. A quantity of electrons of the second plurality of electrons is indicative of the pressure.

Abstract: Provided is a mounting base member used in a pair for mounting an object to be conveyed thereon, the mounting base member including: a body part configured to support the object to be conveyed; and a top plate part that extends in at least a horizontal direction from an upper part of the body part. The top plate part covers an upper surface of the body part, or the upper surface of the body part contacting the conveyed object is flush with an upper surface of an extending portion of the top plate part.

Abstract: A computing system and method for identifying related visual content of a collection of visual media files comprising one or more image files and video files includes monitoring inputs to the computing system, the inputs associated with a user interaction with electronic content using the computer system, identifying a visual media file in the collection of visual media files relevant to the electronic content based on a semantic label assigned to the visual media file by the computing system, creating a representative image of the identified visual media file, and displaying the representative image for selection. The computing system enables a selection of the displayed representative image for association of the identified visual media file with the electronic content.

Abstract: A method, apparatus and system for zero shot object detection includes, in a semantic embedding space having embedded object class labels, training the space by embedding extracted features of bounding boxes and object class labels of labeled bounding boxes of known object classes into the space, determining regions in an image having unknown object classes on which to perform object detection as proposed bounding boxes, extracting features of the proposed bounding boxes, projecting the extracted features of the proposed bounding boxes into the space, computing a similarity measure between the projected features of the proposed bounding boxes and the embedded, extracted features of the bounding boxes of the known object classes in the space, and predicting an object class label for proposed bounding boxes by determining a nearest embedded object class label to the projected features of the proposed bounding boxes in the space based on the similarity measures.

Abstract: A phased array antenna system has at least one trough reflector, each trough reflector having at least one phased array located at a feed point of the reflector, and an array of elements located near to a point equal to one half of a center transmission wavelength. A method of decoding a receive signal includes propagating a transmit signal through a transmit and a receive path of a phased array to generate a coupled signal, digitizing the coupled signal, storing the digitized coupled signal, receiving a signal from a target, and using the digitized coupled signal to decode the signal from the target. A method of modeling the ionosphere includes transmitting measuring pulses from an incoherent scattering radar transmitter, receiving incoherent scatter from the transmitting, and analyzing the incoherent scatter to determine pulse and amplitude of the incoherent scatter to profile electron number density of the ionosphere.

Abstract: An example computing device includes a functional encryption unit configured to generate a master secret key and public key; apply functional encryption using the public key to biometric information of a user to produce functionally encrypted biometric information, the functional encryption is based on an encryption function that encodes the biometric information, a computation engine configured to perform re-enrollment by at least one of 1) retrieving a pre-generated function key from a memory, or 2) retrieving a dynamically generated function key from the one or more storage nodes that can be inaccessible during user authentication, the function key dynamically generated using the master secret key, and applying, using the function key and functionally encrypted biometric information, a decryption operation to generate new helper data, wherein the new helper data is generated as an evaluation of a cryptographic function during the decryption operation without the need to decrypt the biometric information.

Abstract: In an embodiment, the disclosed technologies include automatically recognizing speech content of an audio stream that may contain multiple different classes of speech content, by receiving, by an audio capture device, an audio stream; outputting, by one or more classifiers, in response to an inputting to the one or more classifiers of digital data that has been extracted from the audio stream, score data; where a score of the score data indicates a likelihood that a particular time segment of the audio stream contains speech of a particular class; where the one or more classifiers use one or more machine-learned models that have been trained to recognize audio of one or more particular classes to determine the score data; using a sliding time window process, selecting particular scores from the score data; using the selected particular scores, determining and outputting one or more decisions as to whether one or more particular time segments of the audio stream contain speech of one or more particular classes