Abstract: Provided is a device and a method for route planning. The route planning device (100) may include a data interface (128) coupled to a road and traffic data source (160); a user interface (170) configured to display a map and receive a route planning request from a user, the route planning request including a line of interest on the map; a processor (110) coupled to the data interface (128) and the user interface (170). The processor (110) may be configured to identify the line of interest in response to the route planning request; acquire, via the data interface (128), road and traffic information associated with the line of interest from the road and traffic data source (160); and calculate, based on the acquired road and traffic information, a navigation route that matches or corresponds to the line of interest and meets or satisfies predefined road and traffic constraints.

Abstract: Methods and systems for budgeted and simplified training of deep neural networks (DNNs) are disclosed. In one example, a trainer is to train a DNN using a plurality of training sub-images derived from a down-sampled training image. A tester is to test the trained DNN using a plurality of testing sub-images derived from a down-sampled testing image. In another example, in a recurrent deep Q-network (RDQN) having a local attention mechanism located between a convolutional neural network (CNN) and a long-short time memory (LSTM), a plurality of feature maps are generated by the CNN from an input image. Hard-attention is applied by the local attention mechanism to the generated plurality of feature maps by selecting a subset of the generated feature maps. Soft attention is applied by the local attention mechanism to the selected subset of generated feature maps by providing weights to the selected subset of generated feature maps in obtaining weighted feature maps.

Abstract: The present disclosure is directed to a cleavable agent for enhanced magnetic resonance generally corresponding to the formula Y-L-R, wherein Y represents a catalyst-binding moiety having at least one isotopically labeled heteroatom, L represents a cleavable bond, and R represents a hyperpolarized payload having at least one isotopically labeled carbon. Also disclosed herein is a method of cleaving the cleavable agent for enhanced magnetic resonance.

Abstract: A method and system of image hashing object detection for image processing are provided.

Abstract: Methods and systems for budgeted and simplified training of deep neural networks (DNNs) are disclosed. In one example, a trainer is to train a DNN using a plurality of training sub-images derived from a down-sampled training image. A tester is to test the trained DNN using a plurality of testing sub-images derived from a down-sampled testing image. In another example, in a recurrent deep Q-network (RDQN) having a local attention mechanism located between a convolutional neural network (CNN) and a long-short time memory (LSTM), a plurality of feature maps are generated by the CNN from an input image. Hard-attention is applied by the local attention mechanism to the generated plurality of feature maps by selecting a subset of the generated feature maps. Soft attention is applied by the local attention mechanism to the selected subset of generated feature maps by providing weights to the selected subset of generated feature maps in obtaining weighted feature maps.

Abstract: A sensor calibrator comprising one or more processors configured to receive sensor data representing a calibration pattern detected by a sensor during a period of relative motion between the sensor and the calibration pattern in which the sensor or the calibration pattern move along a linear path of travel; determine a calibration adjustment from the plurality of images; and send a calibration instruction for calibration of the sensor according to the determined calibration adjustment. Alternatively, a sensor calibration detection device, comprising one or more processors, configured to receive first sensor data detected during movement of a first sensor along a route of travel; determine a difference between the first sensor data and stored second sensor data; and if the difference is outside of a predetermined range, switch from a first operational mode to a second operational mode.

Abstract: A method and system of image processing for action classification uses fine-grained motion-attributes.

Abstract: The present disclosure is directed to a cleavable agent for enhanced magnetic resonance generally corresponding to the formula Y-L-R, wherein Y represents a catalyst-binding moiety having at least one isotopically labeled heteroatom, L represents a cleavable bond, and R represents a hyperpolarized payload having at least one isotopically labeled carbon. Also disclosed herein is a method of cleaving the cleavable agent for enhanced magnetic resonance.

Abstract: Provided is a device and a method for route planning. The route planning device (100) may include a data interface (128) coupled to a road and traffic data source (160); a user interface (170) configured to display a map and receive a route planning request from a user, the route planning request including a line of interest on the map; a processor (110) coupled to the data interface (128) and the user interface (170). The processor (110) may be configured to identify the line of interest in response to the route planning request; acquire, via the data interface (128), road and traffic information associated with the line of interest from the road and traffic data source (160); and calculate, based on the acquired road and traffic information, a navigation route that matches or corresponds to the line of interest and meets or satisfies predefined road and traffic constraints.

Abstract: Described herein are advanced artificial intelligence agents for modeling physical interactions. In one embodiment, an apparatus to provide an active artificial intelligence (AI) agent includes at least one database to store physical interaction data and compute cluster coupled to the at least one database. The compute cluster automatically obtains physical interaction data from a data collection module without manual interaction, stores the physical interaction data in the at least one database, and automatically trains diverse sets of machine learning program units to simulate physical interactions with each individual program unit having a different model based on the applied physical interaction data.

Abstract: Methods and systems for budgeted and simplified training of deep neural networks (DNNs) are disclosed. In one example, a trainer is to train a DNN using a plurality of training sub-images derived from a down-sampled training image. A tester is to test the trained DNN using a plurality of testing sub-images derived from a down-sampled testing image. In another example, in a recurrent deep Q-network (RDQN) having a local attention mechanism located between a convolutional neural network (CNN) and a long-short time memory (LSTM), a plurality of feature maps are generated by the CNN from an input image. Hard-attention is applied by the local attention mechanism to the generated plurality of feature maps by selecting a subset of the generated feature maps. Soft attention is applied by the local attention mechanism to the selected subset of generated feature maps by providing weights to the selected subset of generated feature maps in obtaining weighted feature maps.

Abstract: Methods and systems for advanced and augmented training of deep neural networks (DNNs) using synthetic data and innovative generative networks. A method includes training a DNN using synthetic data, training a plurality of DNNs using context data, associating features of the DNNs trained using context data with features of the DNN trained with synthetic data, and generating an augmented DNN using the associated features.

Abstract: Methods and systems for budgeted and simplified training of deep neural networks (DNNs) are disclosed. In one example, a trainer is to train a DNN using a plurality of training sub-images derived from a down-sampled training image. A tester is to test the trained DNN using a plurality of testing sub-images derived from a down-sampled testing image. In another example, in a recurrent deep Q-network (RDQN) having a local attention mechanism located between a convolutional neural network (CNN) and a long-short time memory (LSTM), a plurality of feature maps are generated by the CNN from an input image. Hard-attention is applied by the local attention mechanism to the generated plurality of feature maps by selecting a subset of the generated feature maps. Soft attention is applied by the local attention mechanism to the selected subset of generated feature maps by providing weights to the selected subset of generated feature maps in obtaining weighted feature maps.

Abstract: Described herein are advanced artificial intelligence agents for modeling physical interactions. An apparatus to provide an active artificial intelligence (AI) agent includes at least one database to store physical interaction data and compute cluster coupled to the at least one database. The compute cluster automatically obtains physical interaction data from a data collection module without manual interaction, stores the physical interaction data in the at least one database, and automatically trains diverse sets of machine learning program units to simulate physical interactions with each individual program unit having a different model based on the applied physical interaction data.

Abstract: A guard route security method and a computer readable storage medium are provided. The method includes receiving positioning information of a security guided vehicle and generating a security guided vehicle track; searching a low-position camera and a high-position camera around the security guided vehicle according to the positioning information, and feeding back searching information; and turning on the high-position camera to trace and monitor the security guided vehicle according to the searching information. According to the guard route security method, the security guided vehicle is shot and monitored at a high-position camera monitor to achieve a follow-up control of the security guided vehicle with the high-position cameras, which is convenient for the security guards to quickly learn the surrounding environment and carry out security work for guard route more intuitively and effectively.

Abstract: The present invention provides a camera link method and a computer storage medium. The method includes acquiring a real-time video of a camera, selecting a point in the real-time video as a linked point and acquiring coordinates of the linked point; designating a linked camera; and acquiring a real-time video of the linked camera, and turning on the real-time video of the linked camera or positioning the linked camera to the linked point. According to the camera link method and the computer storage medium of the present invention, linkage among multiple cameras, and automatically linkage to multiple dome cameras can be achieved, and the application range is wide.

Abstract: A video playing method for synchronously displaying AR information includes capturing a video code stream containing AR information by an AR camera; extracting the AR information from the video code stream frame by frame, generating subtitle information during said extraction, and storing the subtitle information as a subtitle file; storing the video code stream after said extraction as a video file; combining the subtitle file with the video file to create a general video file; and parsing and playing the general video file on a third-party player. The video with AR information captured by an AR camera can be parsed by a third-party player and synchronously displayed along with the video play.

Abstract: Described herein are advanced artificial intelligence agents for modeling physical interactions. An apparatus to provide an active artificial intelligence (AI) agent includes at least one database to store physical interaction data and compute cluster coupled to the at least one database. The compute cluster automatically obtains physical interaction data from a data collection module without manual interaction, stores the physical interaction data in the at least one database, and automatically trains diverse sets of machine learning program units to simulate physical interactions with each individual program unit having a different model based on the applied physical interaction data.

Abstract: A video map engine system includes a configuration management client, multiple video equipments, a video access server, an augmented reality processor, and an augmented reality client. The parameters of the video equipments includes azimuth angle P, vertical angle T and zoom factor Z of the video equipment, the augmented reality client is adapted for calculating the location where the augmented reality tag is presented in the real-time video according to the values of P, T, Z and the target location carried by the augmented reality tag, and presenting the augmented reality tag on the corresponding location of the real-time video. Therefore, the real-time video is served as the base map, and the augmented reality tag is presented on the base map, thereby achieving a video map effect.

Abstract: A video map engine system includes a configuration management client, multiple video equipments, a video access server, an augmented reality processor, and an augmented reality client. The parameters of the video equipments includes azimuth angle P, vertical angle T and zoom factor Z of the video equipment, the augmented reality client is adapted for calculating the location where the augmented reality tag is presented in the real-time video according to the values of P, T, Z and the target location carried by the augmented reality tag, and presenting the augmented reality tag on the corresponding location of the real-time video. Therefore, the real-time video is served as the base map, and the augmented reality tag is presented on the base map, thereby achieving a video map effect.