Abstract: The present disclosure provides a method for training an image enhancement model, the image enhancement model includes an enhancement module including convolution branches corresponding to brightness intervals; and the method includes: inputting a sample image to the image enhancement model, and acquire a result image output by the image enhancement model; calculating losses including an image loss of the result image relative to a Ground Truth image, and a first constraint loss of brightness histogram constraint of each of the convolution branches of an image output from each of the convolution branches relative to the Ground Truth image; adjusting the enhancement module according to the losses; and in a case where a training end condition is not met, returning to the operation of inputting the sample image to the image enhancement model. The present disclosure further provides an image enhancement method and a computer-readable medium.

Abstract: A verification method for fast source and path embedded with random authentication is provided. The method includes: generating a corresponding verification structure for an expected path according to a predetermined path strategy, embedding different m pieces of fragment information randomly selected with same possibility from the verification structure for the expected path to a header of a data packet to be transmitted in a data flow, and transmitting the data packet to be transmitted with the embedded fragment information to a next hop of routing node of the expected path, performing a verification on the received data packet by the respective intermediate routing node on the expected path, and forwarding the data packet to the next routing node when the verification passes, performing verification on the received data packet by the data flow destination, performing a parsing verification evaluation on the expected path when the verification passes.

Abstract: A first switch port receives a first control packet, a header having a source media access control (MAC) address of a first MAC of a first interface of a sending device and a payload containing a second MAC of a second interface of the same sending device. The first and second interfaces form a bonded interface for failover purposes on the sending device. A second switch port receives a second control packet, a header having the second MAC as its source MAC and a payload containing the first MAC. The switch associates the first and second MAC addresses as related MAC addresses associated with the bonded interface. The switch facilitates failover between the secure interfaces via multiple uplinks while maintaining the port security on the switch by allowing transmission of data from either the first or second MAC address, associated as related MAC addresses, without triggering a port security violation.

Abstract: The disclosure provides a backdoor attack method and apparatus for a malicious URL detection system. The backdoor attack method includes: obtaining original URL samples of backdoor URL samples to be generated; determining position information of separator slashes in each URL of the original URL samples, and obtaining a position number result by numbering the position information; determining a backdoor attack mode of the malicious URL detection system based on the position number result, and generating the backdoor URL samples corresponding to the original URL samples by using the backdoor attack mode; and constructing a training set based on a preset ratio of the backdoor URL samples, training a neural network model by using the training set, and testing an attack strength of the malicious URL detection system in the backdoor attack mode by using the trained neural network model to obtain a real-time attack success rate.

Abstract: A preparation method for glufosinate or a salt, an enantiomer or a mixture of enantiomers in all ratios thereof, the method being especially suitable for the preparation of glufosinate, and greatly shortening steps in an existing preparation process. Especially in the preparation of L-glufosinate, the product can effectively retain the ee value of the raw materials.

Abstract: An intelligent vehicle capturing apparatus for a high-pier bridge includes: a double-layer metal protective net; and a support unit and an actuation unit connected to the double-layer metal protective net. The double-layer metal protective net is mounted to a backside of the high-pier bridge through the support unit. The actuation unit includes a hydraulic link mechanism. The hydraulic link mechanism is connected to one end of the double-layer metal protective net. The hydraulic link mechanism is moved to unwind the double-layer metal protective net on the backside of the high-pier bridge to capture a vehicle. In this way, the apparatus stops and buffers a vehicle that is about to rush out of a high-pier bridge.

Abstract: According to one embodiment, a method, computer system, and computer program product for predicting the movement intentions of ships is provided. The present invention may include obtaining binocular camera visual data of a body of water; identifying physical objects in the body of water; generating feature point clouds of the physical objects in the body of water; performing instance segmentation on the generated feature point clouds; analyzing the generated feature point clouds; predicting the movement intentions of identified ships in the body of water based on the analyzed feature point clouds; and displaying the predicted movement intentions of the identified ships in the body of water on an augmented reality device.

Abstract: A facial expression-based detection method for deepfake by generative artificial intelligence (AI) constructs an AIR-Face facial dataset for generative AI-created face detection training, and uses an untrained information feature space for real and fake classification. Nearest linear detection is performed in this space to significantly improve the generalization ability of detecting fake images, especially those created by new methods such as diffusion models or autoregressive models. The detection method improves the performance of extracting features of generative AI-created faces through phased trainings, and detects generative AI-created faces through the feature space. Compared with other methods, the detection method scientifically and effectively improves the accuracy of generative AI-created face recognition, and fully mines the potential semantic information of generative AI-created faces through phased trainings.

Abstract: According to an aspect of the present disclosure, a radiation element is provided, comprising: a basic radiation element and one or more bandwidth extension structures; wherein the one or more bandwidth extension structures are mounted on the basic radiation element to extend the operating bandwidth of the basic radiation element. The present disclosure has the following advantages: the radiation element according to the present disclosure has one or more bandwidth extension structures to extend the operating bandwidth of the basic radiation element, such that by combining the plurality of bandwidth extension structures and the basic radiation element, the radiation element may work well at bands beyond its original operating band, which eliminates the need of using a plurality of basic radiation elements due to different operating bandwidths as required, thereby saving costs.

Abstract: A display panel includes a substrate and a plurality of first sub-pixels and a plurality of light-shielding portions which are both located in a functional device configuration region. At least some first sub-pixels each include a first pixel circuit and a first light-emitting device. The plurality of first light-shielding portions are arranged on a side of a plurality of first light-emitting devices away from the substrate at intervals. An orthographic projection of a first portion, whose orthographic projection is located outside an orthographic projection of a first electrode of the first light-emitting device on the substrate, of the first pixel circuit on the substrate is located within an orthographic projection of a first light-shielding portion of the plurality of light-shielding portions on the substrate.

Abstract: A computer-implemented method for verifying ownership of a deep learning (DL) model includes: obtaining a Gaussian mixture distribution for initialization of a DL model to be verified, a model chain to be verified of the DL model, and a corresponding verification dataset, wherein the model chain is composed of model parameters of the DL model in each training epoch; calculating, based on the Gaussian mixture distribution, the model chain and the verification dataset, an accuracy monotonicity index, a parameter distance monotonicity index, a parameter distribution continuity index, an initial parameter distribution index, an initial parameter randomness index and a model chain distance index; and obtaining an ownership verification result by determining an authenticity of the model chain based on the accuracy monotonicity index, the parameter distance monotonicity index, the parameter distribution continuity index, the initial parameter distribution index, the initial parameter randomness index and the model c

Abstract: The present disclosure discloses a video denoising method and device, and a computer-readable storage medium. The video denoising method includes: dividing each video frame in an input video frame sequence into sub-image blocks, and calculating a block variance of each of the sub-image blocks; calculating an average variance of the sub-image blocks in a current video frame according to the calculated block variances, determining a noise level of the current video frame according to the average variance, selecting a denoise strength and a noise characteristic curve that match the noise level; and denoising the current video frame according to the denoise strength and the noise characteristic curve.

Abstract: The present disclosure provides methods and systems of maintaining a map suitable for guiding autonomous driving. In some embodiments, the method may include receiving a sensor dataset acquired by a sensor subsystem, wherein: the sensor dataset includes information about a road, the sensor subsystem comprises multiple different types of sensors; determining, by a processor, a confidence level by comparing the sensor dataset and the map that includes prior information about the road; in response to determining that the confidence level exceeds a confidence threshold, processing the map by the processor; and storing the processed map as an electronic file, wherein the processed map is configured to guide an autonomous vehicle to operate on the road.

Abstract: A display substrate includes a substrate, scan signal lines, a plurality of first pixel circuits and a plurality of first light-emitting devices. The plurality of first pixel circuits are disposed on the substrate, and are located in a functional device setting region. At least one first pixel circuit includes a first transistor. At least one first light-emitting device includes a first electrode. The at least one first light-emitting device includes at least one first color light-emitting device, and the at least one first pixel circuit includes at least one first color pixel circuit, and a first color pixel circuit is electrically connected to a first color light-emitting device. An orthographic projection of a channel region of a first transistor in the first color pixel circuit on the substrate is at least partially overlapped with an orthographic projection of a first electrode of the first color light-emitting device on the substrate.

Abstract: The present disclosure provides a data processing system, method, and device for monitoring a high definition map data collection trip. A current vehicle location is determined. Whether the vehicle is following a planned route linked to a data collection trip, based on the current vehicle location is determined. On a condition that the vehicle is not following the planned route, a reason why the vehicle is not following the planned route is determined. An alert is generated, displayed on an in-vehicle navigation device, and uploaded with the reason to a command center.

Abstract: The present disclosure provides a data processing system, method, and device. A method for monitoring a high definition map data collection trip includes recording data collected from one or more sensors associated with a vehicle, determining a current vehicle location, and determining whether the vehicle is following a planned route based on the current vehicle location. On a condition that the vehicle is not following the planned route, the method includes including an indication in the recorded data collected from the one or more sensors associated with the vehicle that the vehicle is off-route; generating additional navigation instructions, wherein the additional navigation instructions return the vehicle to the planned route; and providing the additional navigation instructions to a driver of the vehicle.

Abstract: This application describes a motor that includes a housing, an output shaft, a conductive bearing, and a conductive component. The output shaft is disposed in the housing. A positioning hole is disposed on an end face of a first end of the output shaft, and the positioning hole is disposed coaxially with the output shaft. The conductive bearing is disposed in the positioning hole, and includes an inner ring, and an outer ring that is fixedly connected to the output shaft. One end of the conductive component is fixedly connected to the inner ring of the conductive bearing, and the other end is connected to the housing. In this application, the motor can output, by using the conductive bearing and the conductive component, charges accumulated in the output shaft, to eliminate a generated common-mode voltage.

Abstract: Methods and systems for providing secure file sharing are described herein. A computing device may receive data from a resource provider in response to a request received by the resource provider. The data may include information about a user of a client device that made the request for a file and information about the file, and the request may be for access to the file and made by the user with an owner of the requested file being unknown to the user. The computing device may parse the received data to identify a subset of data. The subset of data may be configured to enable the owner to determine whether to permit the user to access to the file. Further, the computing device may provide a notification, which includes at least some of the subset of data, to an endpoint device of the owner.

Abstract: A method and system for determining acoustic emission (AE) parameters of rock based on moment tensor analysis. The method includes: constructing, according to macroscopic mechanical parameters, a numerical model of a rock specimen to be tested; loading the numerical model through particle flow code software to simulate a failure process of the rock specimen to be tested, and identifying fracturing time and positions of microcracks when the PFC software loads the numerical model; determining, when the PFC software loads the numerical model, if rock grains of two sequentially generated microcracks include common rock grains, and an interval for generating the two microcracks is less than duration time of a present AE event, the two microcracks as a same AE event; taking geometric centers of all microcracks within a spatial range of an AE event as source positions of the corresponding AE event; and determining AE parameters of the AE event.

Abstract: The present application discloses an autoregression image abnormity detection method of enhancing a latent space based on memory, which belongs to the field of abnormity detection in computer vision. The present application comprises: selecting a training data set; constructing a network structure of an autoregression model of enhancing a latent space based on memory; preprocessing the training data set; initializing the autoregression model of enhancing a latent space based on memory; training the autoregression model of enhancing a latent space based on memory; verifying the model on the selected data set, and using the trained model to judge whether the input image is an abnormal image. In the present application, a prior distribution is not needed to be set such that the distribution of the data itself will not be destroyed, and it can prevent the model from reconstructing abnormal images, and ultimately can better judge abnormal images.