Abstract: The present invention belongs to the technical field of computer vision, and proposes a lightweight real-time emotion analysis method incorporating eye tracking. In the method, gray frames and event frames that have synchronized time are acquired through event-based cameras and respectively input to a frame branch and an event branch; the frame branch extracts spatial features by convolution operations, and the event branch extracts temporal features through conv-SNN blocks; the frame branch has a guide attention mechanism for the event branch; and the spatial features and the temporal features are integrated by fully connected layers, The final output is the average of the n fully connected layer outputs, which represents the final expression. The method can recognize the emotional expression of any stage in various complex light changing scenarios; and in the case of limited accuracy loss, the emotion recognition time is shortened to achieve “real-time” user emotion analysis.

Abstract: A method for preparing a solar cell is provided. The method includes providing a N-type silicon substrate; depositing a tunnel passivation structure on the first surface of the N-type silicon substrate, and then depositing a mask layer on the tunnel passivation structure; cleaning the second surface of the N-type silicon substrate; performing boron diffusion treatment on the cleaned second surface of the N-type silicon substrate and annealing treatment on the tunnel passivation structure in the same environment, so that a first emitter layer is formed on the second surface of the N-type silicon substrate and the tunnel passivation structure is crystallized; performing laser patterning treatment on the first emitter layer to form a second emitter region; depositing a passivation and anti-reflection film; and forming a first electrode and a second electrode.

Abstract: The present disclosure discloses an image encryption method based on multi-scale compressed sensing and a Markov model. According to the difference in information carried by low-frequency coefficients and high-frequency coefficients of an image, different sampling rates are set for the low-frequency coefficients and the high-frequency coefficients of the image, which can effectively improve the reconstruction quality of a decrypted image. The decrypted image obtained by the present disclosure has higher quality than the decrypted image generated by the existing scheme, and a better visual effect and more complete original image information can be obtained.

Abstract: An embedded bionic winch for sampling and exploration of a polar sub-glacial lake, including an actuating chamber, a power chamber, a transition chamber, a sensor chamber, a cable arranging chamber and a slip ring chamber, where the actuating chamber, the power chamber, the transition chamber, the sensor chamber, the cable arranging chamber and the slip ring chamber are arranged in sequence, and are coaxially connected; the power chamber provides a power for the winch, and realizes sealing of motors located therein under water with a certain depth; the transition chamber realizes a power transmission between the power chamber and the cable arranging chamber; the sensor chamber is used for mounting of a tension sensor; the cable arranging chamber realizes retraction and release of a cable through the precise cooperation of a drum and a lead screw, and the slip ring chamber contains a slip ring.

Abstract: The application discloses a signal transmission method and device. The signal transmission method comprises: a first communication node determining a mode for transmitting data in a time unit to a second communication node; and the first communication node transmitting data to the second communication mode in the time unit according to the determined mode for transmitting data. The data includes N data sets, where N is a positive integer. A data transmission method for M data sets in the data is known by the second communication node. Related information of a data transmission method for (N?M) data sets is notified to the second communication node by the first communication node by means of a control signal in the time unit, where M is less than or equal to N.

Abstract: The present application provides a large-sized single crystal cutting device and a method of using the same. The device includes a cutting system, a commutation system, and a detection system. The cutting system includes a cutting line, a pay-off axis, a guide wheel, and a take-up axis. The commutation system is disposed between the pay-off axis and the take-up axis to adjust a tension and a line speed of the cutting line. The detection system is configured to detect the tension and the line speed of the cutting line.

Abstract: A solar cell is provided. The solar cell includes: an N-type silicon substrate having a first surface and a second surface, a tunnel passivation structure and a first passivation and anti-reflection film formed on the first surface, a boron-doped emitter structure layer including a first emitter layer and a second emitter region formed on the second surface, a second passivation and anti-reflection film formed on the emitter structure layer, a first electrode configured to be in electrical contact with the second emitter region, and a second electrode configured to be in electrical contact with the tunnel passivation structure. The solar cell of the present application has a selective emitter structure. The metal contact region has a large junction depth to meet the metallization requirements. The region outside the metal contact region has a small junction depth to improve the optical response.

Abstract: A solar cell is provided. The solar cell includes: an N-type silicon substrate having a first surface and a second surface, a tunnel passivation structure and a first passivation and anti-reflection film formed on the first surface, a boron-doped emitter structure layer including a first emitter layer and a second emitter region formed on the second surface, a second passivation and anti-reflection film formed on the emitter structure layer, a first electrode configured to be in electrical contact with the second emitter region, and a second electrode configured to be in electrical contact with the tunnel passivation structure. The solar cell of the present application has a selective emitter structure. The metal contact region has a large junction depth to meet the metallization requirements. The region outside the metal contact region has a small junction depth to improve the optical response.

Abstract: The present disclosure relates to a tactile pattern Super Resolution (SR) reconstruction method and acquisition system, which belong to the field of tactile perception. First, a High Resolution (HR) tactile pattern sample is obtained by using a Low Resolution (LR) tactile sensor; then, a deep learning-based tactile SR model is trained by using a tactile SR data set; and finally, reconstructing the tactile data of a contact surface to be measured as an SR tactile pattern by using the tactile SR model. The present disclosure uses the existing taxel-based LR tactile sensor and adopts a deep learning-based tactile SR reconstruction technology, which can effectively restore the shape of the contact surface, improves the resolution of the tactile sensor, and meanwhile, maintains the characteristics of the sensor being light, flexible, and easy to be integrated into devices, such as a robot.

Abstract: Example neural architecture search methods and image processing methods and apparatuses in the field of computer vision in the field of artificial intelligence are provided. The example neural architecture search method includes determining search space and a plurality of construction units, superimposing the plurality of construction units to obtain a search network, adjusting, in the search space, network architectures of the construction units in the search network, to obtain optimized construction units, and establishing a target neural network based on the optimized construction units. In each construction unit, some channels of an output feature map of each node are processed by using a to-be-selected operation to obtain a processed feature map, and the processed feature map and a remaining feature map are stitched and then input to a next node.

Abstract: A processor-implemented method for a neural architecture search (NAS) starts by generating an over-parameterized super network having multiple layers. The super network has multiple operator types. Each of the layers includes a largest super kernel corresponding to a search space. The method also includes performing gradient descent to evolve a largest super kernel to a small kernel corresponding to the search space in order to generate a range of kernel encodings. The method further includes identifying a subset of kernel encodings from the range of kernel encodings, for each layer of the super network, based on the gradient descent. The method determines a set of candidate architectures based on the subset of kernel encodings, each of the candidate architectures having a different model size. The method selects a target model, from the set of architectures, based on meeting hardware specifications, and then applies the target model.

Abstract: A method for constructing an artificial intelligence (AI) integrated model is provided, including: obtaining a training dataset, an initial graph network model, and a plurality of base models; then iteratively training the initial graph network model by using training data in the training dataset and the plurality of base models, to obtain a graph network model; and then constructing the AI integrated model based on the graph network model and the plurality of base models, where an input of the graph network model is a graph structure consisting of outputs of the plurality of base models. Since the graph network model considers neighboring nodes of each node in the graph structure when processing the graph structure, the graph network model fully considers differences and correlations between the base models when fusing the outputs of the plurality of base models.

Abstract: An antenna structure, an electronic device, and a wireless network system are provided, and relate to the field of antenna technologies. A patch antenna array includes four patch antennas. The four patch antennas are arranged in two rows and two columns. One of the feeding structures is included between two of the patch antennas in each row. One of the feeding structures is included between two of the patch antennas in each column. The feeding structure located between the two patch antennas in each column is connected to the first feeding port, so that the four patch antennas all generate polarization in a first direction. The feeding structure located between the two patch antennas in each row is connected to the second feeding port, so that the four patch antennas all generate polarization in a second direction.

Abstract: The present invention belongs to the technical field of preparing isocyanate derivatives, and particularly relates to a storage-stable polyisocyanate composition and a preparation method therefor. The method comprises: in the presence of a catalyst system, subjecting an isocyanate monomer to a polymerization reaction; terminating the reaction after a suitable conversion rate is reached, so as to obtain a polyisocyanate reaction solution; and then performing heat treatment on the obtained polyisocyanate reaction solution to obtain a heat-treated polyisocyanate mixture, wherein the heat treatment temperature is 10-30° C. higher than the heat-sensitive temperature of the obtained polyisocyanate product, and the heat treatment time is 5-30 min; and then performing separation treatment on the heat-treated polyisocyanate mixture. When the polyisocyanate composition of the present invention is stored at 50° C.

Abstract: This application relates to the field of image annotation and recognition in the field of artificial intelligence technologies, and in particular, to a data annotation method. The method includes: using at least two different classification models; pretraining one of the classification models as an initial classification model, and annotating a label for data in a to-be-annotated source dataset as initial data by using the pretrained classification model; and controlling the classification models to perform alternating training and data annotation a quantity of times. Operations of current training and current data annotation include: obtaining data that is re-annotated with a label by a previously trained classification model, selecting a first part of the data to train a current classification model, and re-annotating, by the trained current classification model, a label for a second part of data that is not selected.

Abstract: A model training method and a related apparatus are provided, which may be used in computer vision to perform image detection. The method includes: extracting feature information in a target image; further separately extracting features of a target object from the feature information by using a Gaussian mask to obtain a first local feature and a second local feature; determining a feature loss by using the first local feature and the second local feature; performing prediction by using the first network and the second network based on a same region proposal set to obtain a first classification predicted value and a second classification predicted value, and obtaining a classification loss based on the first classification predicted value and the second classification predicted value; and training the second network based on the classification loss and the feature loss to obtain a target network.

Abstract: This application relates to the field of artificial intelligence, and specifically, to the field of computer vision. The method includes: performing feature extraction on an image to obtain a basic feature map of the image; determining a proposal of a region possibly including a pedestrian in the image; processing the basic feature map of the image to obtain an object visibility map in which a response to a pedestrian visible part is greater than a response to a pedestrian blocked part and a background part; performing weighted summation processing on the object visibility map and the basic feature map to obtain an enhanced feature map of the image; and determining, based on the proposal of the image and the enhanced feature map of the image, a bounding box including a pedestrian in the image and a confidence level of the bounding box including the pedestrian in the image.

Abstract: Example neural architecture search methods and image processing methods and apparatuses in the field of computer vision in the field of artificial intelligence are provided. The example neural architecture search method includes determining search space and a plurality of construction units, superimposing the plurality of construction units to obtain a search network, adjusting, in the search space, network architectures of the construction units in the search network, to obtain optimized construction units, and establishing a target neural network based on the optimized construction units. In each construction unit, some channels of an output feature map of each node are processed by using a to-be-selected operation to obtain a processed feature map, and the processed feature map and a remaining feature map are stitched and then input to a next node.

Abstract: In an approach for debugging and simulating application runtime execution, a processor loads source code and logs into a debug tool. A processor generates log debug information including a log map, a log variable cross reference table, and a method call stack tree. A processor determines a plurality of log blocks based on log context in the logs and the method call stack tree. A processor maps the source code to the logs for each log block. A processor suggests a starting point and a breakpoint based on the log variable cross reference table and the log blocks. A processor compares a source code variable value to a log variable value and a source code execution path to a log execution path based on the mapping set between the source code and the logs. A processor simulates the source code variable value with the log variable value.

Abstract: A keyword extraction method is provided. In the method, a candidate keyword set in a target text is obtained by processing circuitry of a server. An extraction degree of the candidate keyword is determined by the processing circuitry based on subject similarity and a text conversion frequency of a candidate keyword in the candidate keyword set. The subject similarity is between the candidate keyword and the target text. The extraction degree indicates a probability at which the candidate keyword used as a keyword matching the target text is extracted. The keyword is extracted by the processing circuitry from the candidate keyword set according to the extraction degree.