91results about How to "Convenient guidance" patented technology

The invention discloses a simulation method for abrasive particle vibration machining based on reaction molecular dynamics, and relates to the technical field of reaction molecular dynamics, a vibration auxiliary model of a Si surface of 4H-SiC is established, the process of vibration auxiliary chemical mechanical polishing of diamond abrasive particles in the Z-axis direction is simulated, the Si surface model of the 4H-SiC is consistent with an actual machining surface of silicon carbide, and the abrasive particle vibration machining simulation method is suitable for the actual machining surface of silicon carbide. The vibration direction of the vibration auxiliary simulation is consistent with the motion direction of the abrasive particles in the actual chemical mechanical polishing process, so that the acting force and motion trail simulation of the diamond abrasive particle model in the polishing process are more real, and the simulation process is closer to the actual machining process; the atom removal mechanism of the Si surface of 4H-SiC under the conditions of different vibration amplitudes and frequencies and different scratching depths can be researched, so that a chemical mechanical polishing technology is conveniently optimized by ultrasonic vibration, and the polishing quality of silicon carbide is further improved; the technical problem that an existing vibration auxiliary model based on reaction molecular dynamics simulation cannot effectively optimize the chemical mechanical polishing process of silicon carbide is solved.

The invention relates to the technical field of pesticide detection, and discloses an analysis method and system of an all-in-one fluorescent quantitative test strip with a self-adaptive threshold adjustment function, the method comprises the following steps: firstly obtaining historical detection data of N fluorescent signal channels of a historical detection sample, and determining a first threshold parameter of each fluorescent signal channel; then acquiring real-time detection data of a to-be-detected sample, calculating first dynamic parameters of the N fluorescence signal channels, and dynamically adjusting the first threshold parameter to obtain a second threshold parameter; a fluorescence characteristic parameter, a sample matrix characteristic parameter and an environment interference parameter are obtained to correct the second threshold parameter, and self-adaptive thresholds of N fluorescence signal channels are obtained; and finally performing quantitative analysis and result evaluation on the to-be-detected sample according to the self-adaptive threshold value. According to the invention, the threshold value can be dynamically adjusted according to the change of the detection environment, and the accuracy and reliability of the detection result of the all-in-one fluorescent quantitative test strip are improved.

The invention relates to a multi-modal data feature layer fused sow weight estimation method, which relates to a detection method, and comprises the following steps: acquiring point clouds of sows at different angles through a sow point cloud acquisition device, and taking sow time sequence data such as variety, fetal number and reproduction stage and reconstructed sow real-time three-dimensional point clouds as input; a dual-branch network fused in a feature layer is established, meanwhile, deep features are automatically extracted from one-dimensional data and three-dimensional point clouds, time sequence data features and real-time data features are fused, and current body weight estimation of the sow and body weight curve estimation of the whole breeding stage are completed. Through the method, the accuracy of weight estimation is improved, the final output of the method is the weight real-time value of the current reproduction stage and the weight prediction value of the whole reproduction stage, and the real-time weight value is beneficial for obtaining the physical health condition of the sow at the current stage and adjusting the feeding strategy; therefore, the sows can reach the optimal health state and maintain the optimal production performance, and the minimum feed cost is maintained.

The invention belongs to the technical field of underground engineering construction, and provides a shield tunneling parameter optimization method and system based on historical data mining, and the technical scheme is as follows: constructing a shield tunnel geologic model based on an obtained geological survey report and advanced geological forecast data; tunneling safety grading is conducted on the shield tunneling surrounding rock based on the shield tunnel geologic model, and a dynamic geologic model is obtained; based on the dynamic geologic model, the site construction environment and related specifications, set shield construction task information is obtained; according to the dynamic geologic model and set shield construction task information, corresponding shield tunneling parameter information is determined, shield tunneling efficiency under the current shield tunneling parameters is analyzed, if the shield tunneling efficiency does not reach the set level, the shield tunneling parameters are adjusted according to the shield tunneling efficiency condition under the current shield tunneling parameters, and the shield tunneling efficiency is adjusted according to the shield tunneling efficiency condition under the current shield tunneling parameters. And guiding field construction according to the optimized shield tunneling parameters. The shield tunneling efficiency is improved, and the engineering cost is reduced.

The invention belongs to the field of artificial intelligence and financial science and technology, and relates to a human face three-dimensional reconstruction method and device based on artificial intelligence, computer equipment and a storage medium, and the method comprises the steps: carrying out the region-of-interest extraction of two-dimensional human face image data, obtaining a human face image, carrying out the key point detection of the human face image, and obtaining two-dimensional key points; performing face segmentation on the face image to obtain a face mask image; processing the face image based on the initial reconstruction model to obtain a three-dimensional face model; generating a first loss based on the three-dimensional face model and the face image; generating a second loss based on the three-dimensional face model, the two-dimensional key points and the face mask image; generating a comprehensive loss based on the first loss and the second loss; optimizing the initial reconstruction model based on the comprehensive loss to obtain a target reconstruction model; and performing three-dimensional reconstruction processing on the face image based on the target reconstruction model to obtain a target three-dimensional face model. According to the invention, the precision of three-dimensional face reconstruction is improved based on the use of the target reconstruction model.

The invention discloses a forest fire danger grade forecasting method based on a subcompartment scale, and belongs to the technical field of forest fire prevention, the forest fire danger grade forecasting formula is P = (T + S) * M, P is a forest fire index; t is a combustible type index; s is a forest fire source index; m is the moisture content index of the fine dead combustible. The problems that existing forest fire risk prediction is low in fineness, low in precision and complex in prediction process are solved.

The invention provides a multi-mode integrated circuit chip performance evaluation system and method, and the method comprises the steps: dynamically controlling the temperature, power supply voltage and clock frequency parameters in a performance evaluation box environment where an integrated circuit chip is located according to the working mode of the integrated circuit chip; the performance evaluation box is used for detecting environmental parameters of the integrated circuit chip, sending the detected environmental parameter data to a performance evaluation management platform, and testing the power consumption, the data transmission rate, the clock precision and the optical transmittance of the integrated circuit chip when the environmental parameters in the performance evaluation box are adjusted; processing and analyzing the output power consumption, the data transmission rate, the clock precision and the optical transmissivity data, generating an evaluation report, sending the evaluation report to the performance evaluation management platform, and displaying the evaluation report to a user by the performance evaluation management platform; the system corresponding to the method comprises an environment parameter adjusting module, a testing module, a processing and analyzing module and a display module, and the performance of the multi-mode integrated circuit chip can be comprehensively and accurately evaluated through the method and the system.

The invention belongs to the technical field of image segmentation, and discloses an improved image segmentation method and system. The invention designs a self-attention mechanism based on Euclidean distance, and puts forward a U-Net image segmentation algorithm based on the Euclidean distance self-attention mechanism by embedding the self-attention mechanism into U-Net. According to the self-attention mechanism based on the Euclidean distance, modeling is conducted on a large-range context information relation by paying attention to the difference degree of information representation between different feature points, then global features are extracted, and the information difference is emphasized. A self-attention mechanism based on the Euclidean distance is introduced into the U-Net, so that the large-range information relation of network modeling can be better guided, and the information difference between feature points is concerned and compared, thereby improving the learning ability of the model to global features and the discrimination ability to similar feature points, making up for the defect of fuzzy segmentation boundary, and further improving the segmentation precision.

The invention discloses a demand optimization-oriented hydrothermal geothermal resource evaluation method, and relates to the technical field of geothermal resource quality and development and utilization time sequence classification evaluation, and the method comprises the steps: determining a relative importance relationship between evaluation indexes of an evaluation object through an expert scoring system, and carrying out the evaluation of the evaluation object based on the relative importance relationship. And constructing a judgment matrix for the evaluation object, determining the weight and the standardized numerical value of each evaluation index, multiplying and accumulating the weight of each evaluation index and the corresponding standardized numerical value to obtain a comprehensive score, and determining an evaluation result of the evaluation object according to the comprehensive score. It can be seen that the relative importance relation of the evaluation indexes is determined through the expert scoring system, the indexes of the resource demand side and the resource supply side are comprehensively considered, the judgment matrix is constructed, accurate evaluation of the evaluation object is achieved, and therefore the actual development decision is better guided.

The invention discloses a low-concentration gas detection method based on a direct absorption spectrum, which comprises the following steps of: calculating SO2 concentration by using real-time measured spectral data of detected gas and a detected gas absorption cross section database measured in a laboratory through an ultraviolet differential absorption spectrometry, calculating to obtain an interference spectrum of SO2 on NO2, removing the influence of SO2 on NO2 absorbance, and determining the concentration of the NO2. The absorbance of the zero absorption position is calculated, the influence of spectral energy reduction on measurement is eliminated, the NO2 concentration after cross interference is eliminated is accurately calculated, and the detection precision and the detection lower limit are greatly improved.

The invention discloses a learning image compression method based on side information autoregression, and the method comprises the steps: taking a natural image x as input, obtaining a compact potential representation y, and directly coding y when the image x is actually compressed; the potential representation y is used as input, side information z is obtained, and the side information z is used for providing prior guidance for entropy coding of the potential representation y; a side information autoregression module is used for estimating probability distribution of side information z, and a compressed bit stream of the side information z is formed after entropy coding; the quantized side information z is sent to a hyper-synthesis transformation module to obtain hyper-prior information potentially representing y; estimating the probability distribution of the potential representation y by using the context of the potential representation y and the hyper-prior information, and forming a compressed bit stream of the potential representation y after entropy coding; and the quantized potential representation y is sent to a synthesis transformation module to obtain a final reconstructed image # imgabs0 #. The reconstructed image obtained by the method is more excellent in the aspect of structural similarity objective indexes, and texture details of the reconstructed image obtained in the aspect of subjective quality are reconstructed better.

The invention relates to the technical field of artificial intelligence, in particular to a self-adaptive website architecture building method based on artificial intelligence driving, which comprises the following steps of: determining a target of a website, a target user group and a demand of the target user group, formulating a project plan, comprehensively considering a plurality of technical stacks according to a plurality of dimensions, and establishing a self-adaptive website architecture. Further selecting an optimal technical stack; according to the selected optimal technology stack, defining an adaptability function to measure the matching degree of existing modules of the website to a user target, defining a performance function to measure the page loading speed, further quantifying the loading performance of the website, and finally evaluating the website architecture characteristics according to the adaptability and the loading performance; according to the method, the accuracy, the foresight and the practicability in the technology stack selection process are improved, the adaptability evaluation is more comprehensive and dynamic, more personalized and immediate response service experience can be provided, and the actual performance of page loading can be reflected more accurately.

The present invention discloses a multi-cell electrothermal coupling simulation method and system for an IGBT chip. The method and system are based on the principle of a unit cell electrothermal coupling model, divide the IGBT chip into detailed partitions, divide the IGBT chip layer by layer based on the particularity of the IGBT structure under electrothermal coupling, and obtain a large-scale IGBT chip electrothermal coupling model by combining and applying multiple partitions to the electrothermal coupling model. The electrothermal coupling simulation of the IGBT chip under the condition of continuous changes in external voltage and current is realized through the multi-cell electrothermal coupling model. Compared with the finite element method, the present invention can obtain the detailed distribution of the surface temperature of a large-area chip changing with time and the detailed situation of the change of the internal current of the chip changing with time more quickly, and accurately simulate the electrothermal coupling characteristics of the chip.

The invention relates to the technical field of information processing, and discloses an ERP (Enterprise Resource Planning) system-based document automatic transmission method, which comprises the following steps of: 1, based on an ERP system, establishing an online data collection module, a user interface and self-service learning module, a user interface calculation module, an ERP system evaluation module and an ERP system management module; step 2, historical user demand collection, new demand collection and receipt type and management problem collection are carried out in an online data collection module; 3, designing an interactive course, a simulation operation and a self-service guide in the user interface and self-service learning module, and designing a plurality of query modes; 4, performing data calculation in a user interface calculation module; 5, the ERP system evaluation module summarizes the results of all the calculation formulas, and comprehensively evaluates the performance and the user satisfaction of the user interface and the self-service learning module; and step 6, the ERP system management module formulates and executes rectification measures according to an evaluation result.

The invention discloses a construction parameter processing method and device and a nonvolatile storage medium. The method comprises the steps that multiple construction parameters corresponding to multiple moments during reservoir transformation construction are obtained, and the multiple construction parameters are used for conducting transformation construction on a reservoir; according to the multiple construction parameters, distribution of fluid in a target shaft at the target moment is determined, the multiple moments comprise the target moment, and the target shaft is used for conveying fluid used for transforming a reservoir to the reservoir; and determining the target bottom hole pressure at the target moment according to the distribution of the fluid in the target shaft. The technical problem that in the prior art, a data analysis method for reservoir transformation construction parameters is not deep is solved.

The utility model belongs to the technical field of laser lithotripsy, and discloses an in-vitro laser lithotripsy efficiency detection device. The device comprises a container, a liquid injection mechanism and a collector, the container is provided with a first opening and a second opening, the first opening is used for inserting a laser fiber, the second opening is located at the bottom of the container, the liquid injection mechanism is communicated with the container and can inject liquid into the container, and the collector is provided with a third opening and at least one fourth opening. The third opening is formed in the top of the collector, the fourth opening is formed in the bottom of the collector, the third opening is communicated with the fourth opening, the second opening is communicated with the third opening, and the size of the fourth opening is not larger than that of the second opening. By means of the device, in-vivo lithotripsy conditions can be simulated, so that detection is closer to clinical practical operation, particles obtained after lithotripsy can be distinguished according to the size, the lithotripsy effect of laser lithotripsy is evaluated, and parameter setting of clinical practical operation can be guided easily.

The present invention discloses a method and device for predicting the activity of multi-view molecule pairs based on twin neural networks, including: constructing a molecule pair activity data set containing a protein target; extracting the initial embedding representations of two molecules and the initial embedding representations of the protein target respectively; constructing a multi-view molecule pair activity prediction model with a twin neural network architecture; optimizing the parameters of the multi-view molecule pair activity prediction model for pre-training tasks and fine-tuning tasks using the activity data of each pair of molecule pairs; and predicting the activity of molecule pairs using the multi-view molecule pair activity prediction model after parameter optimization. The method and device obtain the intramolecular interaction representation and the protein-ligand interaction representation through the multi-view molecule pair activity prediction model, and use the twin neural network to learn the activity differences between molecule pairs, and finally achieve efficient and accurate prediction of molecular activity, so as to better guide the optimization of lead compounds.

In an image projection device configured to project a display image on an image display board, formation of an unnecessary image in a peripheral area of the display image due to generation of stray light is effectively suppressed. A plurality of columnar protrusions 52 is formed on an inner surface 50a of a housing 50 in which an image generation unit that generates the display image is housed, and a second columnar protrusion 54 having a lower height than each of the plurality of the columnar protrusions 52 is formed. That is, even when a part of emitted light from the image generation unit, sunlight, etc. reach the inner surface 50a of the housing 50, the light that has reached the inner surface 50a is reflected multiple times between the plurality of the columnar protrusions 52 and between the plurality of the columnar protrusions 52 and the second columnar protrusions 54, which can effectively attenuate intensity of the light. This reduces a proportion of reflected light and effectively suppresses an occurrence of stray light.

The present invention provides a method, device, terminal and storage medium for judging the economic efficiency of vehicle starting. The method comprises: obtaining key feature data of the vehicle's historical starting state and engine fuel efficiency economic index data, wherein the key feature data is data characterizing the driver's operating characteristics when the vehicle starts; clustering the key feature data and the engine fuel efficiency economic index data to obtain the optimal class in which the engine fuel efficiency economic index data is in the optimal working range of the engine; judging the economy of the vehicle at subsequent starting according to the key feature data corresponding to the optimal class. The present invention can make a more scientific and reasonable judgment on the economy of the vehicle starting and better guide the driver's starting operation.

The invention relates to a knowledge-enhanced big language model reasoning method based on experience knowledge. Comprising the following steps: constructing a knowledge base; obtaining reasoning problems and corresponding problem solving experience; extracting the problem solving experience into structured knowledge entries; a reasoning structure of the problem is extracted in the structure abstraction process, and experience is refined into a reusable solution, judgment and correction mode in the mode generalization process; retrieval of knowledge: for a new problem, related items are retrieved by using a reasoning structure of the new problem; the diversity reordering step then selects a final set of entries, providing a strategically diverse context for reasoning. The problem-based core reasoning structure adopts a diversity perception retrieval mechanism, so that the LLM reasoning accuracy is improved, and the LLM can learn and evolve from practice under the condition that re-training is not needed.

The present invention discloses a vibration pile driving simulation device, comprising: a pile body; an electric vibrator connected to the pile body, suitable for generating an exciting force and transmitting the exciting force to the pile body; and a soil container, suitable for containing the soil body and suitable for vertically upward movement relative to the pile body, so as to simulate the pile driving process of the pile body under the action of the exciting force. The vibration pile driving simulation device of the present invention has a simple structure and can accurately simulate the vibration pile driving process of the pile body under the action of the exciting force in reality.

The invention relates to the technical field of engineering hydraulics, in particular to a curve water flow optimization method based on rough strip flow guide effect evaluation index analysis, which comprises the following steps of: arranging a measurement section and a measurement point at a curve section to obtain water depth, flow velocity and flow data, introducing a water surface transverse gradient discrete coefficient and an average flow velocity discrete coefficient as evaluation indexes, and optimizing the curve water flow. The method comprises the following steps: determining a calculation section, constructing an original data matrix, carrying out forward and standardized processing on a coefficient, calculating an information bearing capacity, determining a weight, calculating a diversion speed reduction weighting coefficient by adopting a multi-index comprehensive weighting method, introducing a dynamic adjustment coefficient, and optimizing rough strip arrangement parameters in combination with a genetic algorithm and a local search strategy. The weighting coefficient adapts to the water flow energy change, the global optimal solution can be efficiently searched, the curve water flow form is improved, the downstream river channel scouring is relieved, and the stability of the release structure and the downstream river channel is improved.

The invention relates to a simulation part design method for simulating a disc forging pre-rotation process, and belongs to the field of aero-engine tests. The method solves at least one problem that the existing pre-rotation process simulation method for the high-temperature alloy large disc forge piece has a large gap between the simulation result and the actual simulation result, the simulation efficiency is low, and the mechanical property change of the pre-rotated material is difficult to accurately evaluate. The simulation part design method comprises the following steps: S1, establishing an initial model; s2, simplifying the model; s3, judging whether linear elastic mechanical analysis results of the initial model and the simplified model are consistent or not; s4, optimizing the rotating speed in linear elastic mechanical analysis, or adjusting the simplified model, and taking the simplified model and the rotating speed which can generate consistent results as a simulation part model and a process test rotating speed; and S5, making a simulation part based on the simulation part model, and carrying out a pre-rotation test at a process test rotating speed. According to the invention, the stress-strain and performance change of the actual disc forge piece in the pre-rotation process can be accurately and efficiently simulated.

The utility model discloses a multi-dimensional load device for muscle-bone ultrasonic evaluation, which comprises a rack, a hip supporting module is fixed on a base of the rack, foot resistance modules are symmetrically arranged on the base, two shoulder load modules are symmetrically arranged at the top of the rack, and the shoulder load modules are symmetrically arranged on the base of the rack. Hand pulling modules are further arranged on the two sides, far away from the center position, of the top of the rack, so that each joint part of the body of a patient in the sitting and standing states achieves a certain motion state under the action of loads, the loads of different parts are combined, the patient can conduct all-around muscle-bone ultrasonic evaluation in the sitting and standing states, and the accuracy of muscle-bone ultrasonic evaluation is improved. As an effective auxiliary means for rehabilitation treatment, by means of load design of buttocks, feet, shoulders and hands, evaluation of muscles, skeletons and related tissues in a motion state can be achieved through ultrasonic real-time monitoring, a doctor can better guide rehabilitation training and treatment selection of a patient, and the rehabilitation effect of the patient is improved. And an effective reference is provided for rehabilitation treatment.

The present invention relates to an image sandstorm removal method integrating color cast correction. The method connects multiple dust perception representation extraction DustAre modules with the same structure, first performs multiple downsampling operations on the input sandstorm image, extracts image features and fuses them; then, performs multiple upsampling operations on the extracted image features and restores them to the original image space to obtain the final sandstorm-removed image. The present invention: (1) In view of the color deviation problem of sandstorm images, the traditional color cast detection and correction strategy is adopted for the label image during network model training, so that the color of the image sandstorm removal result is more natural; (2) Based on the discovery that the chromaticity values ​​of the a* and b* color components of the sandstorm image are much larger than those of the image without sandstorm, a new loss function term is designed to better guide the network model training; (3) A sandstorm image generation model is proposed, and a sandstorm image dataset containing indoor and outdoor labeled images is synthesized for network model training.

The invention belongs to the technical field of tunnel water inrush tests, and particularly relates to a tunnel water inrush test method considering disaster source material deposition-high pressure energy storage. According to the test method, the inclination angle of the launder is adjusted, so that the mud-sand mixture can flow in the launder with different gradients, and the real conditions of different flow speeds of the mud-sand mixture under different geological conditions are simulated. After deposition and solidification of the mud and sand mixture are completed, the deposition box is completely sealed, and water is injected to the set height; after water is injected, a mud-sand mixture, water and gas three-phase mixture in a high-pressure state in the deposition box can simulate a high-pressure energy storage disaster source filled with a solid-liquid-gas three-phase mixture under a real geological condition. According to the test method, test simulation of different geological deposition processes of tunnel filling type karst cave disaster source substances and high-pressure violent water burst catastrophe is implemented, and the test method has good real accuracy.

The invention discloses a retina postoperative recovery device based on facial skeleton features and a control method, the recovery device comprises a chassis, a head tray arranged at the top of the chassis, and a height adjusting assembly arranged between the chassis and the head tray and used for adjusting the height of the head tray; the middle part of the head tray is a head area; personalized fitting is performed according to the facial skeleton features of the patient, and at least three facial skeleton feature points are selected for pressure monitoring, so that the head posture of the patient can be accurately positioned and adjusted, the change of the head posture of the patient can be automatically judged, and a corresponding correct posture prompt is output in combination with the pressure change; the patient is helped to correct bad postures in time, the patient is prevented from keeping the bad postures for a long time, the re-treatment risk caused by improper postures of the patient is reduced, and personal negligence of doctor's advice and self-adjustment of the patient is reduced, so that the postoperative safety of the patient is protected.

The invention relates to a concrete setting state measuring pen and a using method thereof, and belongs to the technical field of concrete testing equipment.The measuring pen comprises a shell, a trocar, a center needle, a balancing weight and an impact spring, the balancing weight is arranged in the shell in a sliding mode, the top end of the trocar is connected with the balancing weight, and the bottom end of the trocar movably penetrates through the bottom of the shell downwards; the impact spring is arranged at the top of the shell, the top end of the center needle is connected with the impact spring, the bottom end of the center needle movably penetrates through the balancing weight downwards and then is connected into the trocar in a penetrating mode, the shell is provided with a side window with a ruler, and the center needle is provided with a marking line matched with the ruler. The insertion depth of the trocar is observed, when the trocar and the center needle are inserted into the concrete, the concrete is in an uncondensed state, when the trocar cannot be completely inserted into the concrete, and the center needle is almost completely inserted into the concrete, the concrete is in an initial condensation state, and when the trocar and the center needle cannot be inserted into the concrete, the concrete is in a final condensation state.

The present invention relates to the field of artificial intelligence technology, and in particular, to a method for building an adaptive website architecture based on artificial intelligence driving, which comprises the following steps: clarifying the goals of a website, a target user group and the needs of the target user group, formulating a project plan, and then comprehensively considering multiple technology stacks according to multiple dimensions, and then selecting the optimal technology stack; based on the selected optimal technology stack, defining an adaptability function to measure the degree of matching of existing modules of the website to user goals, defining a performance function to measure the page loading speed, and then quantifying the loading performance of the website, and finally evaluating the website architecture characteristics according to the adaptability and loading performance; the method improves the accuracy, foresight and practicality of the technology stack selection process, and makes the adaptability evaluation more comprehensive and dynamic, which helps to provide a more personalized and instantly responsive service experience, and can also more accurately reflect the actual performance of page loading.

The present invention discloses a method for secure authentication of vehicle-mounted wireless devices based on deep metric learning, and belongs to the technical field of machine learning. The method comprises: pre-processing radio frequency signals of multiple vehicle-mounted wireless devices to generate a training data set; constructing an open class authentication model based on improved class anchor loss; removing a part of the training data set as improper signal samples, and sending the remaining samples to network training respectively to achieve network model optimization; and authenticating the signals of the vehicle-mounted wireless devices to be authenticated. The present invention achieves efficient identification of vehicle-mounted wireless devices and malicious attack devices while improving the feature clustering effect by constructing an open class recognition model with improved class anchor loss, and balances the recognition rates of the two. At the same time, sparsity is introduced by using a multi-scale feature extraction module, and features are extracted by using different convolution kernels between adjacent convolution layers, and the results of multiple layers of convolution are spliced ​​as the input of the next layer, thereby improving the authentication accuracy of the model without increasing the computational overhead.