Abstract: Methods and systems are provided for automated graphical prescriptions with deep learning systems. In one embodiment, a method for a medical imaging system comprises acquiring, by the medical imaging system, localizer images of a subject, generating, by a trained neural network system, a graphical prescription using the localizer images, and performing, by the medical imaging system, a scan of the subject according to the graphical prescription. In this way, a desired region of interest of the subject may be accurately scanned with minimal input from an operator of the medical imaging system.

Abstract: An electronic apparatus includes an imaging element having a plurality of image capture regions, each of the image capture regions having a plurality of pixels for generating an image signal; a setting unit that sets different image capture conditions for the plurality of image capture regions; and a control unit that corrects a portion of an image signal of a photographic subject captured under first image capture conditions in an image capture region among the plurality of image capture regions so that it is as if the portion of the image signal was captured under second image capture conditions.

Abstract: A volume adjustment method includes performing analysis on the collected sound signal surrounding a user terminal to obtain composition information, where the composition information includes sound types included in the sound signal and proportions of sounds of the various types, and the sound types include blank sound, human sound, and noise, determining a current scene mode of the user terminal according to the composition information, and adjusting volume of the user terminal according to the determined scene mode.

Abstract: Thus, aspects of the present disclosure address model “blow up” by changing the functionality of the activation, thereby providing “dead” or “dying” neurons with the ability to recover from this situation. As one example, for activation functions that have an input region in which the neuron is turned off by a 0 or close to 0 gradient, a training computing system can keep the neuron turned off when the gradient pushes the unit farther into the region (e.g., by applying an update with zero or reduced magnitude). However, if the gradient for the current training example (or batch) attempts to push the unit towards a region in which the neuron is active again, the system can allow for a non-zero gradient (e.g., by applying an update with standard or increased magnitude).

Abstract: A magnetic resonance imaging (MRI) system, comprising: a magnetics system comprising: a B0 magnet configured to provide a B0 field for the MRI system; gradient coils configured to provide gradient fields for the MRI system; and at least one RF coil configured to detect magnetic resonance (MR) signals; and a controller configured to: control the magnetics system to acquire MR spatial frequency data using non-Cartesian sampling; and generate an MR image from the acquired MR spatial frequency data using a neural network model comprising one or more neural network blocks including a first neural network block, wherein the first neural network block is configured to perform data consistency processing using a non-uniform Fourier transformation.

Abstract: Machine learning models can be trained to vend profiles with a high likelihood of matching with a user comprises: training machine learning models with feature vectors representing potential profile matches identified as input to a user match selection process, updating the trained models with feature vectors identified as further input to the selection process, in response to the updating, swapping the trained model out of and a further trained machine learning model into a foreground execution space, resulting in the untrained model being in the foreground and the trained model being in a processing background, after satisfying a defined criterion, injecting an outlier potential match entity as supplemental input to the further training with a supplemental feature vector as further input to the selection process, determining respective confidence values corresponding to candidate profiles accessible to the selection process, and rendering profile images of candidate profiles for the selection process.

Abstract: Devices, systems, and methods for enabling voice communications between a plurality of mobile device users are disclosed. At least one embodiment described herein is directed to a method of establishing a short-range wireless connection for voice communications between multiple mobile devices, where the method comprises detecting one or more second mobile devices within a short-range wireless communication range of a first mobile device, and for each of at least one of the one or more second mobile devices establishing the short-range wireless connection between the first and second mobile devices, where the connection is established automatically in response to the detecting.

Abstract: A style-based generative network architecture enables scale-specific control of synthesized output data, such as images. During training, the style-based generative neural network (generator neural network) includes a mapping network and a synthesis network. During prediction, the mapping network may be omitted, replicated, or evaluated several times. The synthesis network may be used to generate highly varied, high-quality output data with a wide variety of attributes. For example, when used to generate images of people's faces, the attributes that may vary are age, ethnicity, camera viewpoint, pose, face shape, eyeglasses, colors (eyes, hair, etc.), hair style, lighting, background, etc. Depending on the task, generated output data may include images, audio, video, three-dimensional (3D) objects, text, etc.

Abstract: A method for forming a group from a plurality of Bluetooth devices includes forming one subgroup with some of the plurality of Bluetooth devices; and, with respect to Bluetooth mediating devices that are at least some of the plurality of Bluetooth devices belonging to the subgroup, forming, by the Bluetooth mediating device, one lower subgroup having the subgroup as an upper subgroup thereof, together with the plurality of Bluetooth devices which do not yet belong to any subgroup.

Abstract: The present disclosure discloses a Bluetooth mesh network system having power management mechanism that includes low power nodes and a friend node. The friend node operates a part of the low power nodes under a first connection mode to establish a one-to-one connection with the friend node under a connection protocol. The friend node operates the other part of the low power nodes under a second connection mode to establish a many-to-one connection based on a broadcast protocol. The friend node sets a mode operation time threshold value and accumulates a mode connection time for each of the low power nodes. When the mode connection time of one of the low power nodes that operates under one of the first connection mode and the second connection mode reaches the corresponding mode operation time threshold value, the one of the low power nodes switches to operate under the other one of the first connection mode and the second connection mode.

Abstract: An in-vehicle monitoring device includes both forward facing and rear facing cameras. The in-vehicle monitoring device can detect occurrence of one or more triggering events and flag portions of captured videos. The flagged video portions may be transferred to other devices.

Abstract: A method of providing access to IoT resources is provided. The method is performed by a Bluetooth low energy, BLE, tag supporting a Bluetooth wireless technology standard. The method comprises broadcasting a reference to a resource directory comprising reachability information to one or more Internet of Things, IoT, resources and supporting a Constrained Application Protocol, CoAP. A corresponding method performed by a communications device is also provided, and a BLE tag, a communications device, computer programs and computer program products.

Abstract: A radio-frequency impedance tuner can include first and second nodes, a bypass path, first and second series capacitance paths, and an inductance path, with each path being implemented between the first and second nodes and including a switch configured to allow the path to couple or uncouple the first and second nodes. The tuner can further include first and second shunt paths, with each shunt path being implemented between the second node and ground and including a switch configured to allow the shunt path to couple or uncouple the second node and the ground. The tuner can further include a switchable grounding path implemented along the inductance path and configured to allow the inductance path to function as a series inductance path between the first and second nodes, or as a shunt inductance path between the ground and a node along the inductance path.

Abstract: A method for dynamic allocation of communication resources includes (1) determining a risk-return characterization of a plurality of communication resource allocations across a plurality of communication channels in a communication system, (2) selecting a first allocation of the plurality of communication resource allocations from the risk-return characterization according to at least one predetermined criterion, and (3) automatically allocating communication resources among the plurality of communication channels according to the first allocation.

Abstract: A distance sensing system includes a first electronic device and a second electronic device. A second processing unit of the second electronic device sends a wireless signal via a second short-range wireless communication unit. When a first processing unit of the first electronic device receives, via a first short-range wireless communication unit, the wireless signal from the second electronic device, the processing unit generates an estimated distance value indicating an estimated distance between the first electronic device and the second electronic device based on signal strength of the wireless signal. When the first processing unit of the first electronic device determines that the estimated distance value is less than a predetermined distance, the first processing unit outputs a notification via the first output unit, so as to send the notification to a user of the first electronic device when the second electronic device is close to the first electronic device.

Abstract: A waveguide device includes a first electrical conductor including a first electrically conductive surface extending along first and second directions, a second electrical conductor including a second electrically conductive surface opposing the first electrically conductive surface, a waveguide located between the first electrical conductor and the second electrical conductor and extending along the first direction, the waveguide including an electrically-conductive waveguide surface opposing the first electrically conductive surface, and a plurality of electrically-conductive rod rows located on opposite sides of the waveguide, each rod row including a plurality of electrically conductive rods arranged along the first direction. At least one of the first electrical conductor and the second electrical conductor includes at least one hole.

Abstract: A method of training a model includes generating processed data by analyzing an input related to an image using a pre-trained model to detect an object in the input. The method includes generating moving area detection data by analyzing the input to determine whether any areas within the input exhibit movement. The method includes generating temporal inconsistency data by analyzing the processed data to determine whether any areas within the processed data exhibit a temporal inconsistency. The method includes identifying whether the input contains an error candidate based on the moving area detection data and the temporal inconsistency data. The method includes receiving confirmation information regarding whether the error candidate is an undetected object within the input. The method includes updating the pre-trained model in response to receiving the confirmation information indicating that the error candidate is the undetected object within the input.

Abstract: A paging system comprise paging buttons configured to transmit a first signal including a paging request; a server configured to: receive the first signal, identify which of the paging buttons transmit the first signal received by the server, and transmit to at least one portable device a second signal including information on an identifier corresponding to the identified paging button that transmits the first signal, and the paging request; and at least one portable device configured to: display the identifier corresponding to the identified paging button and/or the paging request based on the second signal, and transmit to the server a third signal in response to an acceptance input responded to the paging request. The server transmits to at least one portable device a fourth signal confirming that the portable device transmitting the third signal is matched to the identified paging button in response to the third signal.

Abstract: A method for manufacturing a light-emitting device, comprising: forming, over a substrate, a plurality of multilayered light-emitting structures each including a first electrode, a light-emitting layer, and a second electrode; forming, in the substrate, a plurality of grooves that surround the multilayered light-emitting structures individually; forming, over the substrate, a sealing film that covers the multilayered light-emitting structures and the grooves; and separating the multilayered light-emitting structures from one another after forming the sealing film, by cutting the substrate such that, in each groove, part of the sealing film covering a given inner side surface of the groove remains, the given inner side surface being adjacent to any of the multilayered light-emitting structures.

Abstract: The present disclosure provides a static random access memory (SRAM) cell. The SRAM cell includes a plurality of fin active regions foamed on a semiconductor substrate, wherein the plurality of fin active regions include a pair adjacent fin active regions having a first spacing and a fin active region having a second spacing from adjacent fin active regions, the second spacing being greater than the first spacing; a plurality of fin field-effect transistors (FinFETs) formed on the plurality of fin active regions, wherein the plurality of FinFETs are configured to a first and second inverters cross-coupled for data storage and at least one port for data access; a first contact disposed between the first and second the fin active regions, electrically contacting both of the first and second the fin active regions; and a second contact disposed on and electrically contacting the third fin active region.