17results about How to "Increase training data" patented technology

The invention discloses a rainy day target detection method based on paired image data generation and knowledge distillation, and belongs to the technical field of automatic driving target detection. The method comprises the following steps: acquiring a real image of a road scene and a semantic segmentation mask; inputting the semantic segmentation mask and the weather condition text information into a trained improved diffusion model to obtain a road scene generation image under a rainy day condition; introducing a CoT module and an ODConv module into the target detection model for optimization, and then improving the target detection model by adopting a dynamic weighted aggregation loss function to obtain an optimized target detection model; performing knowledge distillation to obtain a rainy day target detection enhancement model; and performing target detection by using the rainy day target detection enhancement model. The method provided by the invention is excellent in performance in a rainy day scene, has good generalization ability, can provide higher-precision target detection in various complex scenes, and effectively improves the detection accuracy and stability.

The invention aims to provide a named entity recognition method and device based on a large language model, and the method comprises the steps: carrying out the data cleaning of a to-be-trained text, carrying out the data enhancement of the cleaned data, and obtaining target data; the lora structure is fused into a large language model LLaMA model, and a new large language model is obtained; training the new large language model by using target data to obtain a target model; and extracting a target entity from the text by using the target model to complete identification of the named entity. Therefore, the problem of insufficient effect caused by less parameter quantity and knowledge in the traditional method is solved, the training data is increased, the effect of the model is improved, the training speed of the model is improved by adopting lora, and the problem of insufficient training resources is solved.

The invention discloses a fault prediction and simulation method based on digital twinning in a smart factory. The fault prediction and simulation method comprises the following steps: step S1, planning a comprehensive work flow of the smart factory; s2, building a digital twinborn simulation model of the smart factory; s3, training a fault prediction model; and S4, fault prediction analysis and positioning. According to the method, the digital twinborn technology and fault prediction are combined, so that the smart factory can establish the digital twinborn simulation model reflecting the real state of the workshop on the virtual digital level based on the digital twinborn technology; the fault prediction model is used for performing virtual testing on the digital twinborn simulation modeling to predict faults possibly occurring in the working process of the smart factory, so that the working reliability and safety of the smart factory are improved, powerful support is provided for subsequent rapid maintenance of the smart factory, and the training sample of the fault prediction model is increased, so that the fault prediction efficiency is improved. Therefore, the fault prediction capability of the fault prediction model is improved.

The invention discloses a no-reference light field image quality evaluation method based on a visual converter, and the method comprises the steps: extracting a sub-aperture image array of a distorted light field image, and extracting a central sub-aperture image and a sub-aperture image array of which the central angle resolution is a * a; inputting a central sub-aperture image of the distorted light field image into a space blocking module to obtain a distorted light field image space block rich in space information, and inputting a sub-aperture image array of which the central angle resolution is a * a of the distorted light field image into an angle blocking module to obtain a distorted light field image angle block; obtaining multi-layer features on a spatial domain of the distorted light field image, obtaining multi-layer features on an angle domain of the distorted light field image, and splicing the multi-layer features to obtain fusion features rich in spatial features and angle features; and carrying out regression on the fusion features by utilizing a quality score prediction network to obtain a quality score of the distorted light field image, and training the network by utilizing gradient back propagation according to a configured related network model and training parameters.

The invention belongs to the related technical field of power system dispatching, and discloses a solver training method and solving method for a power system security constraint unit commitment model, and the training method comprises the steps: obtaining a plurality of unit commitment models, and generating a plurality of groups of near-optimal solutions for each unit commitment model through a near-optimal solution generation method; constructing a training sample set, wherein each training sample comprises parameter characteristics and a near-optimal solution of the unit commitment model; training the graph neural network to obtain a solver; the near-optimal solution generation method comprises the following steps: adding a near-optimal constraint to a unit commitment model to obtain a near-optimal solution generation model: carrying out randomized disturbance on a coefficient of a target function and adding a diversity constraint to obtain an updated model and solving the updated model, updating a newest solution into a currently obtained solution, and adding the currently obtained solution into a solution set; and repeatedly executing the process until a preset number of near-optimal solutions exist in the solution set. And inputting the parameter characteristics of the to-be-solved unit commitment model into the solver so as to quickly predict the value of the power resource scheduling variable.

The invention relates to the technical field of data processing, and discloses a grounding hanging ring identification method and device based on deep learning, equipment and a medium, and the method comprises the steps: carrying out the model lightweight processing of an initial model through original training data based on deep learning, and obtaining a lightweight grounding hanging ring identification model; performing hybrid enhancement processing on the original training data to obtain enhanced training data; training the lightweight grounding hanging ring identification model by enhancing the training data to obtain a reference grounding hanging ring identification model; performing small target detection pre-training based on the reference grounding hanging ring identification model to obtain a pre-trained grounding hanging ring identification model; and carrying out transfer learning processing on the pre-trained grounding hanging ring identification model to obtain a target grounding hanging ring identification model. By implementing the method provided by the invention, a more accurate target grounding hanging ring identification model can be obtained, a more accurate identification result can be identified according to the target grounding hanging ring identification model, and more accurate grounding hanging ring identification can be realized.

The invention relates to the field of intelligent fault diagnosis, in particular to a nuclear power station rotating machine fault diagnosis method based on small samples and related products, and the method comprises the steps: collecting time domain data and frequency domain data of various fault types, and carrying out the range division; the method comprises the following steps: constructing a total sample set, dividing a training set and a test set, constructing a convolutional neural network, obtaining a nuclear power station rotating machinery fault diagnosis model through training, testing and optimizing iterative training through the test set until optimal convolutional neural network parameters are obtained, and constructing a final nuclear power station rotating machinery fault diagnosis model. Through the method based on small sample learning, the training data is effectively expanded, and the problem that effective diagnosis is difficult to perform under the condition that fault data samples are few in a traditional deep learning model is solved; secondly, the convolutional neural network is adopted to process fault data, so that more accurate and detailed features can be extracted, and the fault diagnosis precision is improved; and through repeated iterative training and model optimization, the stability and high efficiency of the final model are ensured.

The invention relates to the technical field of face key point detection, particularly provides a face key point detection method, a storage medium and intelligent equipment, and aims to solve the problem of poor face key point detection accuracy. In order to achieve the purpose, the face key point detection method provided by the invention comprises the steps of obtaining a face image, and inputting the face image into a trained detection model to obtain a face key point detection result, the detection model comprises a backbone network used for extracting image features, a key point position detection head connected with the backbone network and an auxiliary information detection head connected with the backbone network. The face key point detection is carried out by using the trained detection model, and various features of the face key point detection can be effectively integrated and utilized, so that the accuracy of the face key point detection is improved.

The invention discloses a vector drug package detection method based on a yov8 network, and the method carries out the accurate and rapid detection of a drug package image collected by an industrial camera, provides a high-quality detection result, can achieve the accurate detection of various special-shaped packages, and is very high in robustness. According to the invention, a plurality of data enhancement methods are utilized to preprocess training image samples and construct a training sample set, so that the generalization of an algorithm and the readability of images are improved; according to the provided improved C2F model, multiple gradient flow feature outputs are subjected to multi-dimensional attention processing, classification and regression convolution layers are independently optimized through a decoupling detection head, and an object at any angle can be detected more accurately.

This invention belongs to the field of intelligent agent visual navigation technology. It discloses an image comparison and recognition method and device based on reinforcement learning, including the following steps: (1) designing the target recognition process as an MDP problem; (2) designing a domain randomization method that integrates the reservoir sampling method to sample multiple feature values ​​after randomization of the environmental feature domain; (3) gradually increasing environmental variable factors and task difficulty; (4) using a reward function to evaluate the strategy generated by the agent; (5) designing an experience replay mechanism and setting an experience buffer; (6) using a random sampling method for actions to enable the agent to interact with the environment to obtain a large amount of initial experience, while using a network loss experience replay mechanism to filter the experience in the experience pool for learning; (7) dynamically adjusting the exploratory desire coefficient in the agent's learning process, thereby realizing image comparison and recognition. This invention improves the generalization ability of target recognition.

This invention discloses a product selection method and apparatus based on deep learning, relating to the field of computer technology. One specific implementation of the method includes: collecting user behavior data, the user behavior data including a list of items corresponding to each search term; constructing item pairs based on the items in the item list to obtain a list of item pairs corresponding to each search term; constructing a product selection model based on a deep learning network according to the list of item pairs corresponding to each search term; and performing product selection based on the product selection model. This implementation, by constructing item pairs, building a product selection model based on the constructed item pairs and the deep learning network, and performing product selection based on the selection model, can greatly expand the model training data and improve the coverage and accuracy of the product selection results.

The invention discloses a depression grade distribution generation method, which comprises the following steps that face images are used as training data, each face image is marked as a single-mark BDI-II score which is used as a corresponding description degree, the sum of all the description degrees is equal to 1 and is used for representing the depression severity degree of a human face, the value of the BDI-II score is 0-63, and the value of the BDI-II score is 0-63; depression score distribution of each face image is generated through Gaussian distribution; according to the mapping relation between the depression score and the depression level, a four-dimensional mark distribution is set to describe depression level information of the facial image, each element represents the description degree of the corresponding level, and the mark distribution is depression level distribution. The depression score distribution is converted into grade distribution by means of the relationship between depression scores and grades. In the training process, one face image can be used for real depression level learning and can also be used for learning of other related depression levels, the effect of augmenting training data can be achieved, and the problem of insufficient training data is relieved.

A formula-based numerical reasoning question and answer implementation method, device and medium, the question and answer in the numerical reasoning question and answer application scene are obtained through formula knowledge annotation to obtain formula data set and parameter data set, which are respectively used for training of formula generation model and parameter identification model, for new questions in the question and answer application scene, formula information is generated by the formula generation model, the parameter name in the formula information is input into the parameter identification model, the parameter information of the formula is obtained, the formula information and the parameter information are input into the formula calculator, and the answer of the question is output. The construction of the training data can be automatically carried out, the data migration generalization and data expansion are facilitated; the formula calculator of the calculable unit is introduced in the calculation, the problem that the generative language model is not sensitive to the numerical value is avoided; the numerical reasoning is divided into several subtasks, the requirement for the model is reduced, the obtained answer has good explainability, and the modularization decoupling is favorable for model optimization.

The embodiment of the present application provides an image acquisition device, method and related equipment, and relates to the technical field of image acquisition, so as to solve the problem that the image restoration precision of the current lens-free imaging technology is poor, the array density of the photosensitive device in the photosensitive device is improved, which leads to the increase of the surface area of the photosensitive device, the increase of the manufacturing cost, and the color sensitivity of the small volume target object is not obviously enhanced, thereby affecting the practicability and convenience of the lens-free acquisition technology. The device comprises: at least two photosensitive device layers, wherein at least one of the photosensitive device layers close to the target object is used for light transmission, the photosensitive device layer comprises at least one photosensitive device, and the at least two photosensitive layers are stacked; each photosensitive device layer is used for acquiring a group of image data of the target object.

The application discloses a pathological image color restoration method and a scanner based on deep learning, and the method comprises the following steps: reducing the resolution of input data and target data based on a bilateral grid downsampling technology; training a neural network model for the first time by using the input data and the target data after the resolution is reduced; improving the resolution of the input data and the target data after the resolution is reduced based on a bilateral grid upsampling technology; and training the neural network model for the second time by using the input data and the target data after the resolution is improved, so that the training speed of the neural network model is improved, and the staining effect of the pathological image collected by a digital pathology scanner is improved.

The application relates to the technical field of natural language processing, in particular to a precise question and answer data enhancement method and system based on knowledge graph annotation, which comprises the following steps: firstly, a layered medical knowledge graph containing a type layer and an instance layer is constructed; then, a medical named entity recognition model is used to perform graph annotation on seed question and answer pairs, so that structured graph query modes are generated; next, entity replacement deformation and relationship path deformation are performed on the knowledge graph, so that multiple semantically equivalent derived query modes are generated; then, the natural language generator is used to convert the derived query modes into natural language question and answer pairs; finally, confidence filtering is performed based on attribute range constraints, relationship mutual exclusion constraints and path legality constraints. The application can automatically generate precise question and answer pairs with logical rigor and diverse content under the premise of maintaining medical semantic accuracy, effectively expand medical question and answer training data, and significantly reduce manual annotation costs.