18results about How to "Increase training time" patented technology

The invention provides a Chinese electronic medical record named entity recognition method based on a multi-head attention mechanism and a pulse neural network (SNN), and the method comprises the steps: obtaining a dynamic word vector through a pre-training BERT model, obtaining a text code through a bidirectional long short-term memory network (BiLSTM) model, and screening the features of the code through the SNN network for a named entity in an electronic medical record; the multi-head attention mechanism calculates the dependency relationship between characters, enhances the relevance to the context, and finally outputs a final sequence through a conditional random field (CRF) module to obtain an entity recognition result. According to the method, the time cost is remarkably reduced, the efficiency and precision of identifying the named entities in the medical text are improved, and a more efficient and accurate technical solution is provided for the field of medical information processing.

The invention discloses a sow oestrus behavior recognition method and system based on multi-modal feature fusion, and the method comprises the steps: S1, collecting sow monitoring video and audio data, extracting sow oestrus video segment information, matching audio information of a corresponding time segment according to the sow oestrus video segment information, and carrying out the matching of the audio information of the corresponding time segment; preprocessing the data of the video segment information and the audio information to construct a data set; s2, constructing a deep learning algorithm model for sow oestrus behavior recognition by adopting a multi-modal data fusion method; and S3, obtaining a final state prediction result from a to-be-identified data set through the deep learning algorithm model for sow oestrus behavior identification, and obtaining whether the sow is in an oestrus state or not according to the prediction result. According to the invention, accurate identification of the oestrus behavior of the sow is realized.

The invention discloses a PCCP broken wire signal detection method based on improved YOLOv8, and belongs to the technical field of hydraulic engineering detection. The method comprises the following steps: S1, carrying out a PCCP broken wire prototype test by adopting a distributed optical fiber sensor, collecting sound wave signals generated in the test process, and sorting and screening the collected data to obtain a PCCP broken wire one-dimensional sound wave signal data set; s2, processing and converting the one-dimensional sound wave signal into a two-dimensional spectrogram by adopting synchronous extrusion wavelet transform, labeling, and constructing a data set; s3, training the improved YOLOv8 model by using the data set in the S2, and carrying out broken wire signal identification detection by using the improved YOLOv8 model; performance evaluation is carried out, and the model meeting the evaluation requirement serves as a final target detection model and is used for detecting the wire breakage signal under the PCCP body running state. According to the invention, the problem of insufficient detection and identification of the broken wire signal in the operation process of the PCCP body can be solved, and accurate detection is realized.

The invention discloses a rehabilitation training evaluation method based on dream guidance in combination with myoelectricity monitoring. The method comprises the following steps: collecting and preprocessing electroencephalogram data; analyzing the sleep state and judging whether the sleep state is a deep sleep state; performing stimulation guidance of a deep sleep state; repairing the deep sleep dream; performing real-time myoelectricity monitoring by using myoelectricity monitoring equipment after the patient enters the deep sleep state, and performing motor imagery marking in combination with the electroencephalogram signals; and performing rehabilitation training effect evaluation in combination with the myoelectricity data and the electroencephalogram data. According to the invention, a patient is guided to perform nerve rehabilitation treatment in a mode of monitoring dream by electroencephalogram, so that the sleep time is used for training to improve the training duration, and environmental interference factors can be reduced through deep sleep; the muscle activity effect of the dream guiding part can be judged in an auxiliary mode through the myoelectricity monitoring means, and whether the patient is correctly guided or not is judged in combination with monitoring of the brain electricity on the muscle part.

The invention discloses an analog circuit fault signal processing method based on empirical wavelet transform decomposition, and belongs to the field of analog circuit fault signal processing. An existing analog circuit signal has the problems of nonlinearity and serious fault feature confusion. An analog circuit fault signal processing method based on empirical wavelet transform decomposition comprises the following steps: carrying out adaptive wavelet transform decomposition on a circuit fault signal, calculating an initial demarcation point obtained by a scale space method to obtain a mutual information value of a component, and redividing a frequency spectrum to obtain a new decomposition signal component; performing feature information extraction on the signal reconstructed by selecting the optimal component by using the kurtosis index; and the characteristic parameters are classified by using K-means clustering analysis, so that the diagnosis of the soft fault of the analog circuit is realized. The method reduces the calculation complexity of the whole algorithm, improves the efficiency of the algorithm, and improves the recognition accuracy.

The invention discloses a text-to-image diffusion rapid concept erasing method. The method comprises the following steps: (1) acquiring misleading or false images by using an open source website and preprocessing the misleading or false images; (2) constructing a text-to-image diffusion model: designing a target concept erasing strategy by using a knowledge distillation framework; introducing a calibration set to design other concept retention strategies; (3) generating an image corresponding to the text description; according to the method, the specific concept can be erased efficiently, and meanwhile, the semantic consistency of other related concepts is kept, so that the safety and the reliability of an image generation process are remarkably improved.

The present invention relates to the technical field of fitness equipment, and in particular to a treadmill, comprising a frame and an oxygen production system. The frame comprises a base frame, a front column, a rear column and a shell, a handlebar, a control panel and a liquid crystal display are installed on the front column, a heart rate monitoring piece and an emergency stop button are installed on the handlebar, a lifting frame and a foot are installed at the front end of the base frame and the lower part of the rear end of the base frame respectively, a motor system is installed at the left middle part of the front end of the base frame, a lifting motor is installed at the middle position of the middle part of the front end of the base frame, an oxygen production system is installed at the right middle part of the front end of the base frame, gas generated by the oxygen production system extends to the bottom of the control panel through a conduit attached to the front column, a roller is installed on the base frame, a running belt is sleeved on the roller, and safety pedals are installed on both sides of the running belt on the base frame. The beneficial effects of the present invention are: reasonable structure, providing better exercise experience for athletes, and the oxygen production system can help users extend training time and obtain more training volume.

The invention discloses a lower limb rehabilitation device and method for children with cerebral palsy, and relates to the technical field of medical treatment. Comprising a seat and a pedal mechanism arranged at the front end of the seat, the pedal mechanism comprises a permanent magnet motor, a first pedal and a second pedal, and the first pedal and the second pedal are correspondingly arranged on the two sides of the permanent magnet motor; the system further comprises a host, VR glasses, an upper computer, a permanent magnet motor driver and a detection unit of a pedal mechanism. The detection unit is arranged on the pedal mechanism and is in communication connection with the host; the VR glasses and the upper computer are in communication connection with the host; and the permanent magnet motor driver is arranged on the permanent magnet motor and is connected with the upper computer. Games presented by the VR glasses add fun to rehabilitation, multi-sensory stimulation of the VR glasses promotes plasticity of brain nerves, training time can be effectively prolonged, training frequency and intensity can be improved, a new way is developed for rehabilitation treatment of children with cerebral palsy, and the rehabilitation process of the children with cerebral palsy is powerfully promoted; the structure is simple and easy to realize, and the experience of cerebral palsy children is improved.

The utility model discloses a rehabilitation training device for senile weak patients, and belongs to the field of rehabilitation training. The rehabilitation training device comprises a device base, a pedal is arranged on one side of the upper end of the device base, a bottom tray is arranged on the other side of the upper end of the device base, a back tray is arranged on one side of the bottom tray, and an output motor is arranged at the lower end of the bottom tray. And fan blades are arranged at the position of an output shaft of the output motor. According to the utility model, the problem that sweat discharged by the body skin is continuously adhered between the seat and the contact skin during training is solved, the skin of a user can be prevented from being tightly attached to the back tray for a long time through the support of the first support strip, and the sweat is prevented from being continuously adhered to the surface of the skin; and an air flowing position can be reserved on the support, and air flowing can be realized and the body temperature of the user can be reduced by starting the output motor.

A fire prediction and evaluation method, system, device and medium based on explainable machine learning assistance of multiple types of sensors, uses multiple types of sensors for different scenes, collects and integrates a large amount of fire data sets, covers multiple different fire scenes, namely different time, places and the like, and divides the data sets into a training set and a test set; constructing a convolutional neural network sub-model, and training the model by using a data set to realize universal modeling of fire prediction; a principal component analysis method is used and an entropy weight method is combined to carry out disaster situation evaluation on a place where a fire is predicted to occur, so that accurate universality evaluation on the disaster degree is realized; according to the method, disaster assessment is carried out by combining the convolutional neural network sub-model with the principal component analysis (PCA) and the entropy weight method, fire situations of different types and scales can be accurately analyzed and predicted, the accuracy and operability of disaster assessment are improved, and meanwhile the data processing capacity and the output interpretability of the model are enhanced.

The utility model relates to the technical field of rehabilitation nursing training devices, and discloses a rehabilitation nursing training device which comprises a main body, a servo motor is fixedly connected in the main body, a transmission rod is connected to an output end spline of the servo motor, one end of the transmission rod is fixedly connected with a reciprocating assembly, and an annular plate is fixedly connected to the top end of the main body. The interior of the annular plate is slidably connected with a massage plate, the interior of the massage plate is rotatably connected with a massage roller, the interior of the massage roller is fixedly connected with a spring A, one end of the spring A is fixedly connected with a sliding piece, and the top end of the sliding piece is slidably connected with a massage bead; and a heating plate is fixedly connected into the rehabilitation nursing trainer. According to the rehabilitation nursing trainer, through the arrangement of the reciprocating assembly and the massage plate, in the using process, the arms can be massaged, pain is relieved, and the problem that the recovery is slow due to skeleton movement caused by disordered movement of the arms is solved.

This utility model relates to the field of treadmill technology and discloses an intelligent treadmill with status monitoring function. It includes a base frame, with two support rollers rotatably connected to the upper side of the base frame. A variable-speed motor for driving the support rollers is fixedly installed on the back of the base frame. Running belts are provided on the surfaces of the two support rollers. A support frame is welded to the top of the base frame, and a control panel is provided on the top of the support frame. During training, the user wears a monitoring wristband on their wrist. The device monitors the user's body temperature, heart rate, and blood oxygen concentration. When the blood oxygen concentration drops, a solenoid valve opens, allowing oxygen from the oxygen tank to be ejected from the oxygen outlet through the extension tube, increasing the oxygen concentration at the user's face. This promotes increased blood oxygen levels, allowing for safer extension of training time and providing safety protection for users pushing their physical limits.

The invention belongs to the field of camera relocation in the field of computer three-dimensional vision, and discloses a camera relocation method based on key point guidance. A key point guide scene coordinate generation method based on a matching rate is provided, a key point selection network and a training framework thereof are integrated, and a camera relocation process is optimized. 3D point quality in the point cloud is obtained by evaluating motion structure reconstruction, low-quality key points are filtered and re-projected to a 2D space, and a supervision mask optimization key point selection network is generated. And finally, evaluating the reliability and the accuracy of attitude estimation in combination with the interior point ratio. And when the confidence coefficient of the camera pose estimation is not greater than the threshold value, estimating the six-degree-of-freedom pose of the camera by using the scene coordinate data regressed by the scene coordinate regression network as the final camera pose estimation, thereby improving the precision and generalization ability of the pose estimation. Experimental results verify the effectiveness of the method, and the method is suitable for indoor and outdoor scenes.

The application discloses an antenna substrate vibration base frequency optimization method based on machine learning, and belongs to the technical field of large-area thin-wall plate vibration optimization.The method steps are as follows: step one, identifying an active subspace; step two, adaptively constructing a surrogate model in the active subspace; step three, calling a Kriging surrogate model to estimate a sensitivity index; and step four, optimizing an antenna substrate in a reconstructed design space.The application combines a sensitivity analysis method based on an active subspace and an adaptive sampling strategy and a Bayesian optimization method to establish an antenna substrate vibration base frequency global optimization framework based on machine learning, and can efficiently complete the base frequency global optimization design of a large antenna substrate.The optimization verification on a large-scale antenna substrate with complex constraints shows that the method can overcome the problem of premature convergence of a traditional method, and can obtain a design with a better base frequency under less calculation time, and greatly improves the base frequency optimization efficiency of the antenna substrate.

The invention provides an indoor CSI fingerprint positioning method based on deep learning, and relates to the technical field of indoor wireless positioning. According to the method, abnormal value detection is carried out through an isolated forest algorithm, noise interference is eliminated by combining Hampel filtering and wavelet denoising, dimension reduction is carried out on amplitude and phase data by utilizing PCA, the data quality is effectively improved, and a foundation is laid for subsequent model training. A hybrid model combining a CNN, a BiLSTM and an attention mechanism is adopted, the CNN is responsible for extracting local spatial features, the BiLSTM captures a time sequence dependency relationship, the attention mechanism dynamically focuses key features, advantages are complementary, and the fingerprint classification precision is remarkably improved. And outputting the probability distribution of each reference point through a softmax function, selecting five points with the highest probability, and carrying out weighted average calculation on the coordinates by taking the probabilities of the five points as weights. According to the method, the spatial correlation of adjacent fingerprints is effectively utilized, the positioning result is smoothed, and the higher positioning precision is realized while the fluctuation is reduced.

The invention belongs to the related technical field of mechanical fault diagnosis, and particularly discloses a mechanical composite fault diagnosis method and system based on multi-model collaborative decision, and a medium, and the method comprises the steps: expanding a to-be-diagnosed equipment time sequence signal into Y samples through sliding window sampling; respectively inputting the Y samples into X pre-trained fault detection models to obtain fault detection results; correspondingly converting the fault detection result into a fault label; model fault tags are determined based on the sample number corresponding to the fault tags, and two decision tags are selected from the model fault tags; if the difference between the number of the fault detection models corresponding to the two decision labels is not smaller than a preset threshold value, taking the decision label corresponding to a larger number of the fault detection models as a final fault label; and otherwise, taking the decision tag corresponding to more samples as a final fault tag. According to the invention, multi-model collaborative mechanical composite fault diagnosis and detection are realized, end-to-end diagnosis can be carried out, and the accuracy of fault detection is improved.

The invention discloses a target detection system and method under a special environment. The system includes a teacher network and a target detection network. The target detection network includes a student network and a detection network. The teacher network is used to learn the feature expression of each layer of a clear image under a normal environment in a deep neural network. The target detection network is used to identify the target in the image. The student network is the backbone part of the target detection network, which is used to learn the feature mapping relationship between a fuzzy image under a special environment and a clear image under a normal environment. The detection network is the rest of the target detection network, which is used to output the position and category of the target in the image. The teacher network and the target detection network of the system adopt end-to-end training to obtain a trained target detection initial model. When performing target detection in an image, the teacher network of the target detection initial model is removed to obtain a target detection general model. The target detection general model includes a student network and a detection network, and the target in the image is identified by the target detection general model. The system utilizes the principle of the teacher-student network, and the student network is used as a part of the target detection network. The target detection system only increases the model training time, but no longer needs the teacher network in the reasoning stage, so the reasoning time of the target detection network will not be increased. It can be applied to any target detection network and has a wide range of applications.

The invention discloses a method and a system for wiring in a PCB environment by using a reinforcement learning algorithm to assist an A star algorithm, and relates to the technical field of integrated circuit electronic design automation. The method comprises the following steps: optimizing a wiring path of an A star algorithm in a PCB environment by using a Q-Learning algorithm and a double-attenuation strategy; the A star algorithm comprises a total cost function; in a PCB simulation wiring environment, a Q-Learning algorithm and a double-attenuation strategy assisted A star algorithm are used to solve a wiring path, and a shortest path is obtained. According to the method, the wiring process of the classical A star algorithm is assisted through the reinforcement learning algorithm, the performance is optimized through the double-attenuation strategy, the algorithm training time is shortened while the algorithm routing process is prevented from falling into a local optimal solution, and thus the wiring efficiency is improved.