Abstract: The present invention relates to the technical field of power generation of power systems, and in particular to a power supply and energy storage system using carbon as a medium. The power supply and energy storage system generates energy by reacting carbon dioxide and air, collects carbon monoxide by providing a pressurizing apparatus and a gas storage apparatus, and sends the collected carbon monoxide back to an energy supply apparatus during power consumption peaks, thereby balancing power deficits of loads and renewable energy power generation, considering the stability, reliability, and economic efficiency of the power supply and energy storage system, and improving energy utilization efficiency.

Abstract: This invention provides a method for detecting object, which comprises receiving an image; and executing a deep neural network architecture for the image to obtain one or more object bounding box, wherein the deep neural network architecture comprises a two-dimensional discrete wavelet transform.

Abstract: Embodiments of this application disclose a method and apparatus for training an image region segmentation model, and an image region segmentation method and apparatus. The method includes acquiring a sample image set, and each image of the sample image set having first annotation information; generating graph structure data corresponding to a sample image in the sample image set, the graph structure data comprising multiple nodes, and each node comprising at least one pixel in the sample image; determining second annotation information of each node according to the graph structure data and the first annotation information corresponding to the sample image by using a graph convolutional network model, a granularity of the second annotation information being smaller than a granularity of the first annotation information, the graph convolutional network model being a part of an image segmentation model; and training the image segmentation model according to the second annotation information.

Abstract: The present disclosure provides a real-time dynamic prediction system and method of a three-dimensional shape of a high-pressure jet grouting pile.

Abstract: A qualification rate increasing method for finished products based on trademark parameter detection and control is disclosed. Trademark parameters are detected and controlled at the front end of production. Specifically, according to the folding, gluing, and transportation stages of trademarks in a packaging machine, quality of a trademark gluing position is detected and controlled, trademark indentation stiffness is detected and controlled, and a trademark surface friction coefficient is detected and controlled, to determine the suitability of placing the trademarks on the machine to determine whether the trademarks meet production standards. Only a batch of trademarks of samples whose three detection results are all determined to be qualified can enter the production process, and trademark raw materials that do not meet the production standards are eliminated. Raw materials that may produce defect products are eliminated from the source.

Abstract: A defect rate reduction method for finished products is disclosed. First, size parameters of various types of trademarks, indentations, and printed patterns are entered into a database to generate basic images, a corresponding standard trademark image is collected in real time for each basic image, and edge finding parameter optimization is manually performed, to find edge finding parameters that can be used to accurately extract the trademark edge, the indentation edge, and printed pattern edge on the standard trademark image, to complete the database modeling; sampling detection is performed, edge information is extracted based on a database model, a corresponding distance is calculated based on each extracted edge data information, and the information is compared with internal control to determine surface parameter quality of the sample trademark respectively, it is determined whether production requirements are met based on a determination result, and the unqualified batch of trademarks is eliminated.

Abstract: An integrated circuit die includes a magnetic tunnel junction as a storage element of a MRAM cell. The integrated circuit die includes a top electrode positioned on the magnetic tunnel junction. The integrated circuit die includes a first sidewall spacer laterally surrounding the top electrode. The first sidewall spacer acts as a mask for patterning the magnetic tunnel junction. The integrated circuit die includes a second sidewalls spacer positioned on a lateral surface of the magnetic tunnel junction.

Abstract: A method for object and key-point detection, comprising: receiving an image, executing a deep neural network architecture for the image to obtain one or more object bounding boxes; executing the deep neural network architecture for the one or more object bounding boxes to obtain one or more key-point positions corresponding to the one or more object bounding boxes; and outputting the one or more object bounding boxes and the one or more key-point positions.

Abstract: The present disclosure relate to semiconductor structure that includes a substrate and a memory array. The memory array is spaced over the substrate and has a plurality of rows and a plurality of columns. Further, the memory array comprises a first memory cell and a second memory cell that are adjacent at a common elevation above the substrate. The second memory cell is at an edge of the memory array and separates the first memory cell from the edge, and a top surface of the first memory cell is recessed relative to a top surface of the second memory cell.

Abstract: A system for location-based player feedback for video games may include a data collection module, a pattern recognition module, a localization module and a feedback module. The collection module collects gameplay data for a video game. The pattern recognition module analyzes the collected gameplay data to identify a pattern associated with player difficulty. The localization module associates a game world location with the identified pattern. The feedback module presents a message to players at the game world location associated with the identified pattern requesting feedback. The data collection, pattern recognition, and localization modules may include neural networks trained with machine learning algorithms.

Abstract: The present disclosure provides a multifunctional climbing operation platform and an operation method thereof. The multifunctional climbing operation platform includes a main machine, clamping devices, a bolt retightening device, and an auxiliary sleeve replacement box, wherein the clamping devices each include: a fixed seat; and a pair of claw bars swingably arranged on the fixed seat and each including a rear bar body and a front bar body hinged together, wherein when the pair of claw bars are opened, an inner angle between the rear bar body and the front bar body is greater than 180°, the main machine and the clamping devices can climb along an object to be climbed, the bolt retightening device can replace different sleeves from the auxiliary sleeve replacement box, and can screw a bolt on the object. The present disclosure has the advantages of a large avoidance range and diverse functions.

Abstract: This disclosure relates to methods, devices and systems for real-time recognition of restoration of spontaneous circulation (ROSC) in cardio-pulmonary resuscitation (CPR) process. Recognition mechanisms in both time domain and frequency domain are provided for the ROSC recognition, where the time-domain recognition logic may detect the ROSC by recognizing envelope features of sampled signals in the time domain, and the frequency-domain recognition logic may detect the ROSC by recognizing spectral peaks at different frequency points continuously or significant variations of amplitude of spectral peaks in the frequency spectrum.

Abstract: This application provides an output electrode base, a battery module, and an electric vehicle, and relates to the technical field of batteries. The output electrode base provided in this application includes: a base body and a connecting member. A cavity is made in the base body. At least a part of the connecting member is located inside the cavity and fixed to the base body. A plurality of first stop blocks are disposed around an outer peripheral wall of the connecting member, and the plurality of first stop blocks are configured to limit movement of the connecting member. The battery module provided in this application includes: a battery cell; an output electrode base provided in this application; and an output electrode connecting plate, fixed to the output electrode base.

Abstract: This disclosure discloses a microscope system, a smart medical device, an automatic focusing method, and a storage medium. The smart medical device includes an objective lens, a beam splitter, an image projector assembly, a camera assembly, and a focusing device. The objective lens includes a first end and a second end, and the first end faces a to-be-observed sample. The beam splitter is disposed on the second end. The image projector assembly is in communication with the beam splitter, the image projector assembly includes a first lens and an image projection device, and light generated by the image projector assembly enters the beam splitter through the first lens. The camera assembly includes a camera. The focusing device is disposed on the camera assembly, and the focusing device is configured to perform focus adjustment on the camera.

Abstract: A block erase method for a flash memory is provided. The block erase method is to perform block erase on a block with a predetermined block size. The block erase method includes: performing an erase verification on bytes byte-by-byte in the block when performing the block erase; checking an erase step of the byte when the byte does not pass the erase verification; when the erase step of the byte exceeds a predetermined threshold value, performing the block erase with a partitioned block smaller than the predetermined block size, and returning to an erase verification stage to perform the erase verification; and when the erase step of the bytes does not exceed the predetermined threshold value, continuing to perform the block erase with the predetermined block size, and returning to the erasure verification stage to continue to perform the erase verification.

Abstract: This disclosure provides a method for determining visual and auditory attentiveness of vehicle driver, which comprises: determining a visual attentiveness of a driver according to an image captured by a camera installed inside the vehicle; determining a recognition attentiveness of the driver according to sounds obtained by a microphone installed inside the vehicle; deciding whether to issue a reminder to the driver based on the visual attentiveness and the recognition attentiveness of the driver; when it is necessary to remind the driver after determining the driver's visual and recognition attentiveness, one or a combination of reminder steps will be executed; the reminder steps comprising: issuing a visual reminder by a display device in the vehicle; and issuing a auditory reminder by a speaker in the vehicle.

Abstract: A road boundary detection device is a road boundary detection device that acquires a set of lines corresponding to a road boundary from point cloud data as road boundary information. The road boundary detection device includes: a candidate point detection unit that detects each point of road boundary candidates corresponding to candidates of a road boundary from the point cloud data; a candidate point clustering unit that clusters each point of the road boundary candidates; an adjacent cluster reduction unit that reduces a cluster from a distribution of points in clusters in an adjacency relationship by using a predetermined cluster reduction method; a line fitting unit that fits one or more straight lines or curved lines to one or more of the clusters and output fitted lines as road boundary candidates; a line connecting unit that connects some of the fitted lines by using a predetermined analysis method; and an information output unit that outputs a calculated line as the road boundary information.

Abstract: An active clamp flyback circuit includes: a clamp capacitor that is connected to a primary-side winding of a transformer and that is configured to absorb leakage inductance energy of the primary-side winding; an auxiliary switching transistor that is configured to control the clamp capacitor to perform reverse excitation power charging on the primary-side winding by using the auxiliary switching transistor; a first diode, where the first diode is connected in series between the clamp capacitor and the auxiliary switching transistor; and a second diode, where the second diode is connected between the first diode and the clamp capacitor, and the second diode is connected in series between the clamp capacitor and a primary-side auxiliary winding.