Abstract: Disclosed is a switching power supply circuit and a control circuit and a control method thereof. The control circuit comprises a control unit and a driving unit. The control unit controls the driving unit to provide a driving signal with different driving strengths according to a comparison result between a switching voltage and a preset voltage, so as to adjust a speed of turning the power switch off. By controlling the driving strength of the driving signal, a rising rate of the switching voltage can be improved, a spike of the switching voltage can be reduced, various components in the circuit can be protected from being damaged by the spike, and service life and reliability of the circuit can be improved.

Abstract: A silicone pressure sensitive adhesive with amphiphilic copolymers for maintaining adhesion in a moist environment. The amphiphilic copolymers for silicone adhesives include at least one silicone moiety and at least one hydrophilic segment. Such adhesives are applicable to securing medical devices to human skin.

Abstract: An adaptive filter system and a method for controlling the adaptive filter system are described herein. The system can includes one or more filters to attenuate incoming light. The one or more filters can be moved by one or more actuators. The method can capture image data from an imaging device through the one or more filters. Information can be determined from the captured image data. The one or more filters can be moved to a position for capturing image data based on the information.

Abstract: A computer-implemented method and a system for training a computer-based autonomous driving model used for an autonomous driving operation by an autonomous vehicle are described. The method includes: creating time-dependent three-dimensional (3D) traffic environment data using at least one of real traffic element data and simulated traffic element data; creating simulated time-dependent 3D traffic environmental data by applying a time-dependent 3D generic adversarial network (GAN) model to the created time-dependent 3D traffic environment data; and training a computer-based autonomous driving model using the simulated time-dependent 3D traffic environmental data.

Abstract: A coated separator or coated porous membrane comprising a coating on one or two sides of a separator membrane or porous membrane is disclosed. The coating may contain at least one of an adhesion agent, a shutdown agent, and a binder. The coating containing these components does not contain any inorganic or organic heat resistant materials, including ceramic materials, or it contains inorganic or organic heat resistant materials, including ceramic materials, in small amounts. The separator may be a separator that does not, by itself, have shutdown capability. For example, the separator membrane of the separator may be a monolayer separator membrane made of polypropylene. Also disclosed is a battery cell, a secondary battery, and a capacitor containing at least one of the coated separators disclosed herein.

Abstract: A decentralized method for generation and management of health monitoring related tasks in a hyperconverged infrastructure (HCl) environment is provided. The hosts in the HCl environment each include a health agent and a task manager. The health agent collects health results from health checks and stores the health results in a shared database that is shared by the hosts. The task manager generates a health monitoring related task in response to the health results being indicative of a change in health status, and stores the health monitoring related task in a task pool that is also shared by the hosts. Any of the hosts can obtain and execute the health monitoring related tasks in the task pool based on a task priority and load balancing criteria.

Abstract: Systems, methods, and non-transitory computer-readable media are provided for implementing a tracking camera system onboard an autonomous vehicle. Coordinate data of an object can be received. The tracking camera system actuates, based on the coordinate data, to a position such that the object is in view of the tracking camera system. Vehicle operation data of the autonomous vehicle can be received. The position of the tracking camera system can be adjusted, based on the vehicle operation data, such that the object remains in view of the tracking camera system while the autonomous vehicle is in motion. A focus of the tracking camera system can be adjusted to bring the object in focus. The tracking camera system captures image data corresponding to the object.

Abstract: An approach is provided for dynamically identifying a danger zone for a predicted traffic accident and generating a warning about the accident. From a monitoring device at a first location, information about traffic conditions for vehicles at time T0 and at least one subsequent time T0+t is received. Based on speeds of the vehicles and a fixed distance from the first location, a response time is calculated. Within the response time, a dual artificial intelligence (AI) model using shallow and deep learners applied on historical accident data is employed. Based on the dual AI model and the traffic conditions, an accident involving vehicle(s) in the traffic is predicted to occur in a zone. A warning about the accident is generated and sent to device(s) in the vehicle(s). The warning is presented to driver(s) of the vehicle(s).

Abstract: Systems and methods for gesture recognition based on depth information from a camera include, at an electronic device having a camera system, capturing a video frame and depth information associated with the video frame, identifying a foreground portion of the video frame based on the depth information, and determining whether the foreground portion matches a predefined gesture in a database. In accordance with a determination that the foreground portion matches a predefined gesture in the database, the device determines whether one or more subsequent video frames matches the one or more predefined gestures to produce a recognized gesture.

Abstract: A method for employing facial information in unsupervised person re-identification is presented. The method includes extracting, by a body feature extractor, body features from a first data stream, extracting, by a head feature extractor, head features from a second data stream, outputting a body descriptor vector from the body feature extractor, outputting a head descriptor vector from the head feature extractor, and concatenating the body descriptor vector and the head descriptor vector to enable a model to generate a descriptor vector.

Abstract: A method for training a model for face recognition is provided. The method forward trains a training batch of samples to form a face recognition model w(t), and calculates sample weights for the batch. The method obtains a training batch gradient with respect to model weights thereof and updates, using the gradient, the model w(t) to a face recognition model what(t). The method forwards a validation batch of samples to the face recognition model what(t). The method obtains a validation batch gradient, and updates, using the validation batch gradient and what(t), a sample-level importance weight of samples in the training batch to obtain an updated sample-level importance weight. The method obtains a training batch upgraded gradient based on the updated sample-level importance weight of the training batch samples, and updates, using the upgraded gradient, the model w(t) to a trained model w(t+1) corresponding to a next iteration.

Abstract: Video methods and systems include extracting features of a first modality and a second modality from a labeled first training dataset in a first domain and an unlabeled second training dataset in a second domain. A video analysis model is trained using contrastive learning on the extracted features, including optimization of a loss function that includes a cross-domain regularization part and a cross-modality regularization part.

Abstract: A method for improving face recognition from unseen domains by learning semantically meaningful representations is presented. The method includes obtaining face images with associated identities from a plurality of datasets, randomly selecting two datasets of the plurality of datasets to train a model, sampling batch face images and their corresponding labels, sampling triplet samples including one anchor face image, a sample face image from a same identity, and a sample face image from a different identity than that of the one anchor face image, performing a forward pass by using the samples of the selected two datasets, finding representations of the face images by using a backbone convolutional neural network (CNN), generating covariances from the representations of the face images and the backbone CNN, the covariances made in different spaces by using positive pairs and negative pairs, and employing the covariances to compute a cross-domain similarity loss function.

Abstract: A system included and a computer-implemented method performed in an autonomous-driving vehicle are described. The system performs: detecting one or more movable traffic objects; determining one or more target movable traffic objects from the one or more detected movable traffic objects; determining a type of the one or more target movable traffic objects and an traffic object that has a right of way (ROW) in a situation involving the autonomous-driving vehicle. The system further performs: determining a manner of generating a vehicle behavior notification to the target movable traffic object based on the type of the one or more target movable traffic objects and the ROW; and causing a vehicle behavior notification of the determined manner to be generated to the one or more target movable traffic objects.

Abstract: System and method for automatic remediation for a distributed file system uses a file system (FS) remediation module running in a cluster management server and FS remediation agents running in a cluster of host computers. The FS remediation module monitors the cluster of host computers for related events. When a first file system service (FSS)-impacting event is detected, a cluster-level remediation action is executed at the cluster management server by the FS remediation module in response to the detected first FSS-impacting event. When a second FSS-impacting event is detected, a host-level remediation action is executed at one or more of the host computers in the cluster by the FS remediation agents in response to the detected second FSS-impacting event.

Abstract: An action recognition system is provided that includes a device configured to capture a video sequence formed from a set of unlabeled testing video frames. The system further includes a processor configured to pre-train a recognition engine formed from a reference set of CNNs on a still image domain that includes labeled training still image frames. The processor adapts the recognition engine to a video domain to form an adapted engine, by applying non-reference CNNs to domains that include the still image and video domains and a degraded image domain that includes labeled synthetically degraded versions of the frames in the still image domain. The video domain includes random unlabeled training video frames. The processor recognizes, using the adapted engine, an action performed by at least one object in the sequence, and controls a device to perform a response action in response to an action type of the action.

Abstract: An adaptive filter system and a method for controlling the adaptive filter system are described herein. The system can includes one or more filters to attenuate incoming light. The one or more filters can be moved by one or more actuators. The method can capture image data from an imaging device through the one or more filters. Information can be determined from the captured image data. The one or more filters can be moved to a position for capturing image data based on the information.

Abstract: A sensor enclosure comprising a domed cover and a base. The base can be encased by the domed cover. The base comprises an inner frame, an outer frame, one or more wipers, and a powertrain. The inner frame can provide surfaces for one or more sensors. The outer frame, disposed underneath the inner frame, the outer frame includes a slewing ring. The slewing ring comprises an inner ring to which the domed cover is attached and an outer ring attached to the outer frame. The one or more wipers extends vertically from the outer frame, each wiper having a first end attached to the outer frame and a second end attached to a support ring, and each wiper making a contact with the dome cover. The powertrain, disposed within the outer frame, configured to rotate the ring and the dome cover attached to the inner ring.

Abstract: A method for multi-task learning via gradient split for rich human analysis is presented. The method includes extracting images from training data having a plurality of datasets, each dataset associated with one task, feeding the training data into a neural network model including a feature extractor and task-specific heads, wherein the feature extractor has a feature extractor shared component and a feature extractor task-specific component, dividing filters of deeper layers of convolutional layers of the feature extractor into N groups, N being a number of tasks, assigning one task to each group of the N groups, and manipulating gradients so that each task loss updates only one subset of filters.