Abstract: Methods and systems for object detection include training dataset-specific object detectors using respective annotated datasets, each of the annotated datasets including annotations for a respective set of one or more object classes. The annotated datasets are cross-annotated using the dataset-specific object detectors. A unified object detector is trained, using the cross-annotated datasets, to detect all of the object classes of the annotated datasets. Objects are detected in an input image using the unified object detector.

Abstract: A tremor suppression device includes a garment wearable on an anatomical region and including electrodes contacting the anatomical region when the garment is worn on the anatomical region, and an electronic controller configured to: detect electromyography (EMG) signals as a function of anatomical location and time using the electrodes; identify tremors as a function of anatomical location and time based on the EMG signals; and apply neuromuscular electrical stimulation (NMES) at one or more anatomical locations as a function of time using the electrodes to suppress the identified tremors.

Abstract: A portable and wearable hand-grasp neuro-orthosis is configured for use in a home environment to restore volitionally controlled grasp functions for a subject with a cervical spinal cord injury (SCI). The neuro-orthosis may include: a wearable sleeve with electrodes; electronics for operating the wearable sleeve to perform functional electrical stimulation (FES) and electromyography (EMG), the electronics configured for mounting on a wheelchair; and a controller configured for mounting on a wheelchair. The controller controls the electronics to read EMG via the sleeve, decode the read EMG to determine an intent of the user, and operate the electronics to apply FES via the sleeve to implement the intent of the user. The neuro-orthosis may restore hand function. The controller may include a display arranged to be viewed by the subject, for example mounted on an articulated arm attached to the wheelchair.

Abstract: A method is provided for Cross Video Temporal Difference (CVTD) learning. The method adapts a source domain video to a target domain video using a CVTD loss. The source domain video is annotated, and the target domain video is unannotated. The CVTD loss is computed by quantizing clips derived from the source and target domain videos by dividing the source domain video into source domain clips and the target domain video into target domain clips. The CVTD loss is further computed by sampling two clips from each of the source domain clips and the target domain clips to obtain four sampled clips including a first source domain clip, a second source domain clip, a first target domain clip, and a second target domain clip. The CVTD loss is computed as |(second source domain clip?first source domain clip)?(second target domain clip?first target domain clip)|.

Abstract: Methods and systems for training a neural network include generate an image of a mask. A copy of an image is generated from an original set of training data. The copy is altered to add the image of a mask to a face detected within the copy. An augmented set of training data is generated that includes the original set of training data and the altered copy. A neural network model is trained to recognize masked faces using the augmented set of training data.

Abstract: A tremor suppression device includes a garment wearable on an anatomical region and including electrodes contacting the anatomical region when the garment is worn on the anatomical region, and an electronic controller configured to: detect electromyography (EMG) signals as a function of anatomical location and time using the electrodes; identify tremors as a function of anatomical location and time based on the EMG signals; and apply neuromuscular electrical stimulation (NMES) at one or more anatomical locations as a function of time using the electrodes to suppress the identified tremors.

Abstract: Systems and methods for automatically authenticating an incoming call are disclosed. In one implementation a method for automatically authenticating an incoming call includes receiving a call from a calling device. The call includes an identifier associated with the calling device. The method further includes receiving, separately from the call, authentication data associated with a device or a user, determining, using the identifier and the authentication data, that the authentication data is associated with the same calling device that initiated the call, verifying the authentication data, and based on a result of the verification, determining that the call is initiated by an authenticated device or user.

Abstract: A system and method for provides unverified users an ability to act upon private records known to them while protecting user privacy by not reflecting private information back to the unverified user. As an unverified user inputs information related to their identity into an interface, the system searches an indexed database which may include both registered users and/or unregistered customers indexed from a single data source or from disparate data sources.

Abstract: An activity assistance system includes a video camera arranged to acquire video of a person performing an activity, an output device configured to output human-perceptible prompts, and an electronic processor programmed to execute an activity script. The script comprises a sequence of steps choreographing the activity. The execution of each step includes presenting a prompt via the output device and detecting an event or sequence of events subsequent to the presenting of the prompt. Each event is detected by performing object detection on the video to detect one or more objects depicted in the video and applying one or more object-oriented image analysis functions to detect a spatial or temporal arrangement of one or more of the detected objects. Each event detection triggers an action comprising at least one of presenting a prompt via the output device and and/or going to another step of the activity script.

Abstract: Various aspects are disclosed for distributed application management using an embedded persistent queue framework. In some aspects, task execution data is monitored from a plurality of task execution engines. A task request is identified. The task request can include a task and a Boolean predicate for task assignment. The task is assigned to a task execution engine embedded in a distributed application process if the Boolean predicate is true, and a capacity of the task execution engine is sufficient to execute the task. The task is enqueued in a persistent queue. The task is retrieved from the persistent queue and executed.

Abstract: Embodiments of the disclosure provide a system and method of allocating a resource based on myriad input data. In some embodiments, the myriad input data include membership information, claims data, transactional data, etc. The myriad input data are sorted and organized in a meaningful association relationship before applied to a resource allocation modeling algorithm. The resource allocation modeling algorithm provides estimated resource necessary for the application chosen. For example, an insurance company may use membership information, claims data, transactional data, etc., to estimate how much reserves or funds it should hold to cover future claims within a certain timeframe.

Abstract: Systems and methods for providing called devices with sets of context data associated with communication sessions are disclosed. In one implementation, a method for generating context data associated with a communications session may include receiving, from a calling device at a first subsystem, a request to establish a communications session. The request may include a first identifier associated with the calling device. The method may further include receiving, at a second subsystem, activities data associated with the calling device that transmitted the request to establish the communications session. The activities data may include a second identifier associated with the calling device and may be indicative of device activities of the calling device.

Abstract: An activity assistance system includes a video camera arranged to acquire video of a person performing an activity, an output device configured to output human-perceptible prompts, and an electronic processor programmed to execute an activity script. The script comprises a sequence of steps choreographing the activity. The execution of each step includes presenting a prompt via the output device and detecting an event or sequence of events subsequent to the presenting of the prompt. Each event is detected by performing object detection on the video to detect one or more objects depicted in the video and applying one or more object-oriented image analysis functions to detect a spatial or temporal arrangement of one or more of the detected objects. Each event detection triggers an action comprising at least one of presenting a prompt via the output device and and/or going to another step of the activity script.

Abstract: A system and method for provides unverified users an ability to act upon private records known to them while protecting user privacy by not reflecting private information back to the unverified user. As an unverified user inputs information related to their identity into an interface, the system searches an indexed database which may include both registered users and/or unregistered customers indexed from a single data source or from disparate data sources.

Abstract: Methods and systems for object detection include training dataset-specific object detectors using respective annotated datasets, each of the annotated datasets including annotations for a respective set of one or more object classes. The annotated datasets are cross-annotated using the dataset-specific object detectors. A unified object detector is trained, using the cross-annotated datasets, to detect all of the object classes of the annotated datasets. Objects are detected in an input image using the unified object detector.

Abstract: Using various embodiments, methods and systems to provide safeguard a neighborhood or community are described. In one embodiment, a system is configured to receiving a license plate data of a vehicle, fetch indices of a set of known license plate data from a schema object, where the set includes vehicle license plate information of a plurality of vehicles of a neighborhood or community. In one embodiment, the set can be represented in memory using by a binary tree in memory. The system is then configured to sort the binary tree using a tree sorting algorithm. Thereafter, a tree searching algorithm is used to determine whether the license plate data is found in the binary tree. When the license plate data is found in the binary tree, it is determined that the vehicle is known to the neighborhood or community, otherwise it is determined to be not known.

Abstract: An activity assistance system includes a video camera arranged to acquire video of a person performing an activity, an output device configured to output human-perceptible prompts, and an electronic processor programmed to execute an activity script. The script comprises a sequence of steps choreographing the activity. The execution of each step includes presenting a prompt via the output device and detecting an event or sequence of events subsequent to the presenting of the prompt. Each event is detected by performing object detection on the video to detect one or more objects depicted in the video and applying one or more object-oriented image analysis functions to detect a spatial or temporal arrangement of one or more of the detected objects. Each event detection triggers an action comprising at least one of presenting a prompt via the output device and and/or going to another step of the activity script.

Abstract: A method for performing video domain adaptation for human action recognition is presented. The method includes using annotated source data from a source video and unannotated target data from a target video in an unsupervised domain adaptation setting, identifying and aligning discriminative clips in the source and target videos via an attention mechanism, and learning spatial-background invariant human action representations by employing a self-supervised clip order prediction loss for both the annotated source data and the unannotated target data.

Abstract: Embodiments of the disclosure provide a system and method of allocating a resource based on myriad input data. In some embodiments, the myriad input data include membership information, claims data, transactional data, etc. The myriad input data are sorted and organized in a meaningful association relationship before applied to a resource allocation modeling algorithm. The resource allocation modeling algorithm provides estimated resource necessary for the application chosen. For example, an insurance company may use membership information, claims data, transactional data, etc., to estimate how much reserves or funds it should hold to cover future claims within a certain timeframe.

Abstract: The present disclosure relates generally to systems, methods, and devices for interpreting neural signals to determine a desired movement of a target, transmitting electrical signals to the target, and dynamically monitoring subsequent neural signals or movement of the target to change the signal being delivered if necessary, so that the desired movement is achieved. In particular, the neural signals are decoded using a feature extractor, decoder(s) and a body state observer to determine the electrical signals that should be sent.