3356 results about "Sample image" patented technology

Provided is a method for training a scene recognition model. The scene recognition model includes a core feature extraction layer, a global information feature extraction layer, an LCS module of at least one level with an attention mechanism, and a fully-connected decision layer. The method includes: acquiring parameters of the core feature extraction layer and the global information feature extraction layer by training based on a first scene label of a sample image and a standard cross-entropy loss; training a weight parameter of the LCS module of each level, based on a loss value acquired by performing a pixel-by-pixel calculation on a feature map output from the LCS module of each level and the first scene label of the sample image; and acquiring a parameter of the fully-connected decision layer by training based on the first scene label of the sample image and the standard cross-entropy loss.

The invention relates to the technical field of industrial defect detection, and discloses a memory matching industrial defect detection method based on adaptive feature fusion, and the method comprises the steps: generating a plurality of unknown defect samples through an enhanced image-level defect simulation strategy, and helping a model to effectively distinguish a normal mode from an abnormal mode; the method comprises the following steps: extracting multi-scale features of a to-be-detected sample image, a normal sample image and a defect sample image, designing an adaptive hierarchical memory bank architecture, embedding the multi-scale features into a grid memory bank in a layered manner, further improving the reasoning speed through an adaptive core set sampling strategy, and combining multi-scale feature fusion and a defect positioning optimization technology to obtain a defect positioning algorithm. And the difference between the multi-scale features and the normal mode is fully utilized, so that the accuracy of defect detection and positioning is remarkably improved. Compared with the prior art, the defect detection precision and the positioning accuracy can be improved, and meanwhile, the real-time detection requirement can be met.

Provided in the present application are a model training method, a defect detection method and related apparatuses. The embodiments of the present application can be applied to various scenarios such as computer vision. The model training method trains an initial detection model by means of using a first sampled image set containing some of images that have been used for training and a full newly-added second training image set, and thus, compared with using full historical training data and full newly-added training data to train an initial model, more saves time and reduces the GPU hour consumption; adaptive evaluation of corresponding first weights of model parameters restricts updating of the model parameters with respect to historical training data, so as to solve the problem of knowledge forgetting caused by only using full newly-added data for model fine-tuning, thus improving the learning capability of the detection model; and an optimized detection model obtained by using the model training method is used to detect defects in an image for detection, thus improving the effect and accuracy of defect recognition.

The invention relates to the technical field of image segmentation, in particular to a plant salt tolerance response modeling prediction method and system based on a time sequence image. The method comprises the following steps: preprocessing acquired plant sample image data; performing image segmentation on the preprocessed image data by using a plant semantic segmentation model based on U-Net; a plant salt tolerance response prediction model based on TimeSform is constructed; and predicting the plant salt tolerance response grade by using the plant salt tolerance response prediction model. According to the time sequence image-based plant salt tolerance response prediction modeling method provided by the invention, a prediction process integrating image acquisition, preprocessing, dynamic feature extraction and depth time sequence modeling is constructed, so that the efficiency, precision and automation level of plant salt tolerance phenotype recognition are remarkably improved.

The invention provides a desertification monitoring grading and vegetation extraction method and system, and relates to the field of desertification remote sensing monitoring and ecological assessment, and the method comprises the steps: selecting a multi-temporal satellite image and an unmanned aerial vehicle image which cover a full research region according to a preset condition, and carrying out the data preprocessing of the multi-temporal satellite image and the unmanned aerial vehicle image; the method comprises the following steps: constructing a desertification difference index by fitting a feature space of a vegetation index and a surface albedo by using a multi-temporal satellite image, grading the desertification degree of a research area to obtain a grading result, and locking a key monitoring area in the research area according to the grading result; a vegetation sample image of a key monitoring area is obtained from an unmanned aerial vehicle image, HSL color space conversion and hue optimization processing are carried out on the vegetation sample image, and a normalized vegetation index based on HSL is constructed, so that vegetation information of the key monitoring area is finely extracted, a key desertification disaster area is effectively positioned, and the accuracy of the desertification disaster area is improved. And vegetation information in the region is finely extracted.

The invention discloses a CLIP-based double-prompt optimized few-sample industrial anomaly detection method. The method comprises the following steps: respectively constructing a learnable normal text prompt template and an abnormal text prompt template by utilizing learnable normal word vectors and abnormal word vectors; performing image enhancement processing and image abnormal synthesis processing on the normal sample image for model training to generate an enhanced image and an abnormal synthesis image; inputting the learnable normal text prompt template and the learnable abnormal text prompt template into a text encoder in the CLIP model to obtain text prompt embedding; respectively inputting the normal sample image, the enhanced image and the abnormal synthetic image into an image encoder in the CLIP model to obtain visual embedding; and the similarity between text prompt embedding and visual embedding is measured to carry out optimization training on learnable text prompts, so that the trained CLIP model is utilized to carry out industrial anomaly detection.

The invention belongs to the technical field of mechanical equipment wear state monitoring. The method comprises the following steps of: extracting wear rate characteristics, wear degree characteristics, cutting index characteristics, fatigue index characteristics and oxidation index characteristics in an abrasive particle sample image, and respectively clustering to obtain five groups of discrete stage labels; carrying out normalization processing on the five groups of discrete stage labels to obtain a normalized data matrix; according to the data matrix, weights corresponding to all the features are calculated in multiple modes, and the average value of the weights calculated in the multiple modes serves as the final weight of all the features; according to the final weight of each feature and the normalized data matrix, calculating a comprehensive score sequence of all the abrasive particle sample images; performing clustering according to the comprehensive score sequence to obtain a plurality of optimal clustering centers, and determining a current wear stage according to the optimal clustering centers; and more comprehensive and more accurate division of the wear stages is realized.

Disclosed in embodiments of the present application are a multi-modal retrieval model generation method and apparatus, a device, and a storage medium. The embodiments of the present application can be applied to various scenarios such as artificial intelligence, intelligent transportation, and assisted driving. The multi-modal retrieval model generation method comprises: by means of a prefix vector module, performing feature recognition on first sample image modal data to obtain a first image prefix vector; by means of a modal identification module and on the basis of the first image prefix vector, the first sample image modal data, and first sample text modal data, generating a first retrieval character; by means of a constrained decoding module, acquiring from a pre-generated database a first sample retrieval result associated with the first retrieval character; and on the basis of a first reference retrieval result and the first sample retrieval result, adjusting the model parameters corresponding to the prefix vector module and the constrained decoding module to obtain a retrieval model after the adjustment. The present application can improve the training efficiency of multi-modal retrieval models.

The invention discloses a Gaussian splash model training method and device, a medium and a program product, and relates to the technical field of computer graphics, and the method comprises the steps: carrying out the initialization of an original Gaussian splash model based on a sparse input sample image; according to the sparse input sample image and a preset Gaussian densification strategy, multiple rounds of iterative training are carried out on the original Gaussian splash model to obtain a progressive Gaussian splash model, and the Gaussian densification strategy comprises the following steps: when the number of times of iterative training of the original Gaussian splash model meets a preset number range, the number of times of iterative training of the original Gaussian splash model meets the preset number range; the densification threshold value used by the original Gaussian splash model in the iterative training process is inversely proportional to the number of iterative training times. According to the method, a relatively high densification threshold value is used at the initial stage of training, so that an over-fitting risk caused by over-high densification degree at the initial stage of training under a sparse input view angle is reduced.

The invention relates to a costume design system and method based on multi-modal AIGC and a storage medium thereof. Comprising a multi-source input processing module used for receiving and processing multiple modal inputs including costume design text description, a costume style drawing, a fabric sample image and user preference data; the clothing feature extraction module is used for extracting clothing structure features and visual style features from the multi-source input; the semantic understanding module is used for carrying out semantic understanding and style analysis on the clothing features; the multi-modal fusion module is used for mapping the clothing feature representations of different modals to a unified semantic space and generating a fusion feature vector; and the style migration module is used for realizing design generation of a specific style based on a small number of garment samples through a LoRA low-rank adaptation technology, and is used for generating a garment design drawing through a multi-stage conditional diffusion model based on the fusion feature vector.

The invention provides a corn leaf disease and insect pest detection method and device, and relates to the technical field of crossing of agricultural disease and insect pest monitoring and target detection, and the method comprises the following steps: obtaining a disease and insect pest sample image, preprocessing the obtained disease and insect pest sample image, and obtaining image data with labels; dividing the image data with the labels into a training set, a test set and a verification set according to a sample distribution proportion; inputting the training set into a corn leaf disease recognition model to train the model, setting a loss function to evaluate the trained model, setting a sample distribution strategy, and dynamically adjusting a sample distribution proportion according to an evaluation result to obtain an optimal corn leaf disease recognition model; and using the optimal corn leaf disease identification model to carry out disease and pest detection on the corn leaf image. By adopting the corn leaf disease and insect pest detection method and the corn leaf disease and insect pest detection device, the position information and the category information of typical corn leaf diseases and insect pests can be accurately classified and positioned.

The embodiment of the invention provides a model training method and device based on visual reinforcement learning, equipment and a medium. Comprising the following steps: acquiring a sample image frame and semantic category information, inputting the semantic category information into a visual large language model to obtain a first convolution kernel parameter, and inputting the sample image frame into a first feature convolution kernel to obtain a first feature thermodynamic diagram; obtaining a second convolution kernel parameter and a second feature thermodynamic diagram of the sample image frame through a preset visual reinforcement learning model; constructing a first distillation loss based on the first convolution kernel parameter and the second convolution kernel parameter, and constructing a second distillation loss based on the first characteristic thermodynamic diagram and the second characteristic thermodynamic diagram; through prediction and calculation of the sample action data and the sample state data, self-supervision loss and target strategy loss are constructed; and based on the first distillation loss, the second distillation loss, the self-supervision loss and the target strategy loss, performing parameter adjustment on a preset visual reinforcement learning model to obtain a target visual reinforcement learning model.

The invention relates to the field of seepage analysis, and particularly discloses a loess slope seepage three-dimensional observation device and method under circulating seepage, and the method comprises the steps: constructing a loess seepage slope physical simulation experiment model, collecting an experiment sample image by using an image collection device, and analyzing the attribute parameters of the device to determine whether to adjust the device; preprocessing, bilateral filtering and temperature difference analysis are carried out on the collected experiment image, preprocessing process parameters are monitored, preprocessing precision factors are evaluated according to the parameters, and whether secondary processing is carried out or not and temperature abnormal areas are separated or not is judged; matching a temperature mapping parameter set based on the precision factor, evaluating the mapping accuracy, and correcting as required; finally, an initial three-dimensional temperature field is generated in combination with sensor data, and a permeation path is generated for three-dimensional rendering; according to the method, through multi-step cooperation, three-dimensional observation and analysis of loess slope seepage are achieved, a systematic technical scheme is provided for research of loess slope seepage characteristics, and improvement of the accuracy and the visualization level of slope seepage research is facilitated.

The invention relates to the technical field of microscopic biological observation, in particular to a microscopic biological observation method based on a digital microscope. Firstly, original point cloud of a sample area on an objective table is obtained through structured light scanning, noise points are removed, point cloud density is optimized by adjusting scanning intensity for multiple times, high-precision and low-noise preprocessed point cloud is generated, then a reference plane is fitted in combination with the structured light point cloud, and obstacle and non-obstacle features are extracted. Detecting a second center coordinate of the bounding box by using a YOLO model, obtaining a third center coordinate through Hough circle detection, generating a standard coordinate by fusing multi-source coordinates, generating a global path based on a grid map and an RRT algorithm, predicting a track in real time, calculating a minimum Euclidean distance with an obstacle, dynamically inserting a B spline transit point to avoid a dangerous area, and ensuring a safe distance. Finally, collecting sample images in real time along the path and extracting microorganism characteristics; the technical problem that the efficiency is low when a traditional microscope is used for acquiring a microorganism image in a sample is solved.

The invention discloses a fluorescence lifetime imaging phase analysis method based on deep learning, a terminal and a readable storage medium, and the method comprises the steps: obtaining counting data of time-dependent single photon counting of each pixel in a measured sample image, and measuring a response function of a used instrument, inputting the convolution attenuation data and the response function into a deconvolution neural network; the deconvolution neural network outputs a deconvolution signal of each pixel according to the response function and the counting data of each pixel; and performing phase transformation according to the deconvolution signal of each pixel, and obtaining the fluorescence lifetime of each pixel. According to the method, the time migration caused by the instrument response is regarded as the convolution caused by the instrument response function, and then the neural network is adopted to carry out deconvolution on the counting data, so that the time migration caused by the response function is calibrated and corrected, a traditional phase calibration process is not needed, and accurate analysis of the service life is realized.

The invention relates to a few-sample image classification method based on hyperbolic space image-text local feature alignment, and belongs to the technical field of image recognition and artificial intelligence, and the method comprises the steps: generating word-level attribute description for a support set image through employing a multi-mode large language model; encoding the image and the text by adopting a vision-language model; constructing a hyperbolic local feature alignment module in a hyperbolic space, screening most relevant image local features for text local features by calculating hyperbolic cosine similarity, and fusing by using hyperbolic weighted average; designing a hyperbolic cross attention module, and aggregating key information from the multi-modal local features of the support set to construct a category prototype by taking query image aggregation features as guidance; and finally performing classification based on the hyperbolic geodesic distance. According to the method, the hierarchical modeling capability of the hyperbolic space and the semantic priori knowledge of the large language model are fully utilized, fine-grained multi-modal feature alignment is realized, and the small sample image classification performance is improved.

The invention discloses a battery device measuring system, method and equipment and a storage medium, the measuring system comprises a shooting piece, a measuring piece and a battery component conveying line, and the shooting piece and the measuring piece are both arranged on the battery component conveying line. According to the scheme provided by the invention, whether the difference value between the actual pixel row number and the theoretical pixel row number is in the preset range or not is compared, if so, the sample image size is determined to be accurate, and if not, the sample image size is compensated, so that when the preset range is exceeded, the sample image size is not compensated. By compensating the size of the sample image, inaccurate measurement of the size of the battery component caused by slipping of the battery component or losing of the shooting piece is avoided, so that the size and the length of the battery component can be accurately measured.

The invention discloses a form category determination method and device for a power system, and the method comprises the following steps: obtaining different categories of form sample images, and extracting keywords and the positions of the keywords in the form sample images; constructing a relation direction matrix of the category form; obtaining a text block set of the to-be-classified form; constructing a relation direction matrix of the to-be-classified form based on the relative position relation between the text blocks; based on a similarity calculation scheme, calculating a similarity score between the relation direction matrix of the to-be-classified form and the relation direction matrix of each category of form; and thus, the category of the to-be-classified form is judged. According to the method, high robustness can be still kept when the forms shift or the fields change, and the classification accuracy of the forms with complex structures or sparse keywords is remarkably improved; meanwhile, the calculation amount is greatly reduced, a large-scale form library can be adapted, and the application requirement of an actual service scene of a power system is met.

A semantic segmentation model training method and apparatus, an electronic device and a storage medium are provided. The semantic segmentation model training method includes: acquiring a sample image, and extracting visual image features corresponding to the sample image by a semantic segmentation model to be trained; processing the sample image to obtain a text image feature corresponding to the sample image, the text image feature being an image feature generated from language description text for the sample image; fusing the visual image features with the text image feature to obtain multimodal features, and performing image segmentation prediction based on the multimodal features to obtain a target loss; and training the semantic segmentation model to be trained based on the target loss to obtain a target semantic segmentation model.

The invention relates to the technical field of image generation, and discloses a country style and appearance design image generation method and system based on a knowledge graph and a large model, and the method comprises the steps: constructing a core element covering country style and appearance design and a knowledge graph; constructing a multi-level label system based on the knowledge graph; pre-processed rural style and appearance sample images are obtained, the rural style and appearance sample images are labeled based on a multi-level label system, and a special training set in the field of rural style and appearance design is obtained; constructing a basic diffusion large model architecture, and training the basic diffusion large model architecture by using the special training set for the country style and appearance design field to obtain a trained country style and appearance design image generation model; and generating a rural style and appearance design image by using the trained rural style and appearance design image generation model. According to the method, the problems that a general large model is insufficient in domain knowledge support and insufficient in data labeling specialty in rural style and appearance design image generation are solved, and the rural style and appearance design image generation efficiency and quality are improved.

The embodiment of the invention provides a training method and device of an open vocabulary detection model and electronic equipment. The open vocabulary detection model training method comprises the steps of obtaining sample data; the sample data comprises a sample image, a sample text and label information, inputting the sample data into a neural network model to be trained, and extracting text feature information of the sample text and image feature information of the sample image; generating dynamic text prompt information according to the image feature information; performing fusion processing on the text feature information and the dynamic text prompt information to obtain semantic fusion feature information; predicting an object in the sample image according to the semantic fusion feature information; and performing iterative training on the neural network model according to the object prediction result, the label information, the text feature information, the dynamic text prompt information and the semantic fusion feature information to obtain an open vocabulary detection model. According to the invention, the context sensing, semantic generalization and fine-grained understanding capabilities of the open vocabulary detection model can be improved.

A semantic segmentation model training method and apparatus, an electronic device and a storage medium are provided. The semantic segmentation model training method includes: acquiring a teacher semantic segmentation model that is pre-trained, the teacher semantic segmentation model including a first teacher network and a second teacher network, the first teacher network having structural characteristics of low depth and high width, and the second teacher network having structural characteristics of high depth and low width; processing a sample image based on the teacher semantic segmentation model to obtain a first segmentation map and a second segmentation map; and training a student semantic segmentation model that is lightweight according to the sample image, the first segmentation map and the second segmentation map, so as to obtain a target semantic segmentation model.

The invention discloses a transformer fault diagnosis method, system and equipment and a medium, and the method comprises the steps: through the linkage design of an oil pumping sampling module and a transparent sampling channel, combining high-speed camera shooting collection and synchronous light source illumination, and obtaining the image of particles in a transformer oil sample in real time; secondary pollution and personal errors possibly caused by traditional off-line sampling are avoided, and the detection efficiency and the objectivity of data are improved; the ResNet neural network is introduced to carry out feature extraction on the oil sample image, the particle number and particle size distribution information are automatically identified, the complex process of manually designing features in a traditional image processing method is avoided, and the accuracy and robustness of diagnosis are greatly improved; by setting judgment threshold values for indexes such as the particle number, the average particle size and the large particle size proportion, logic judgment is conducted in combination with the actual operation state, potential abrasion, pollution and other abnormalities can be recognized in the early stage, the capacity of quick response and advanced intervention is achieved, and the operation safety and the intelligent operation and maintenance level of the transformer are remarkably improved.

A text-to-image model training method, apparatus, and computer-readable storage medium for enhancing text-to-image generation through object-aware training. The method trains a text-to-image model using cyclic iterative training with sample image and text pairs. Training involves selecting image-text sample pairs containing multiple objects, obtaining corresponding mask images and object class names that distinguish location regions of the objects, and inputting both the sample image with description text and the mask images with object class names into the model. The method obtains image predicted noise and object predicted noises, constructs a loss function based on these predictions, and performs parameter adjustment accordingly. This approach enables improved object-level understanding in text-to-image generation models.

The invention is suitable for the technical field of defect detection, and provides a sample generation method and device based on defect detection and a storage medium, the method comprises the following steps: obtaining an original image and semantic indication information of a target product, the semantic indication information being used for indicating a defect type; based on the original image and the target detection model, determining a detection area in the original image and generating a detection area image; based on the detection area image, determining a target area in the detection area and generating a target area mask image; and generating a defect sample image of the target product based on the detection area image, the target area mask image, the semantic indication information and a defect sample generation model. According to the technical scheme provided by the embodiment of the invention, the requirement of a user on defect generation and the target area for defect generation are fully considered, and the accuracy of the defect sample image is improved; and parameter tuning is not needed in the process of generating the defect sample image, so that the generation efficiency of the defect sample is improved.

The invention relates to a generation method and device of a defective image, electronic equipment and a storage medium, and the method comprises the steps: obtaining a defective sample image, a defect-free sample image, a sample control condition and a sample instruction, the sample control condition is used for indicating a defect background, a defect position and a defect form in the defective sample image, and the sample control condition is used for indicating the defect background, the defect position and the defect form in the defect-free sample image; the sample instruction is used for describing a main category and a sub-category of the defect; sequentially performing coarse training and fine training on the initial image generation model based on the defective sample graph, the defect-free sample graph, the sample control condition and the sample instruction to obtain a trained image generation model; and inputting the defect-free image, the target control condition and the target instruction into the trained image generation model to obtain an output image with defects. According to the invention, the generation efficiency of the defect graph is improved.

The invention provides a multi-class anomaly detection method and system based on a pre-training visual language model in a training data scarcity scene, and relates to the technical field of anomaly detection. A first pre-training visual language model is used for obtaining feature representation of a small number of normal sample images in a text space; using a second pre-training visual language model to obtain global features and block features of a small number of normal sample images, constructing and training an adaptive prompt vector generator, in the training process, updating parameters of the adaptive prompt vector generator to obtain a trained adaptive prompt vector generator, and obtaining a training result of the adaptive prompt vector generator; according to the method, the ability of a pre-training visual language model is effectively combined, a training prompt vector generator is guided, a self-adaptive prompt vector for an anomaly detection task is generated, a one-to-many visual memory warehouse and a one-to-many prompt vector warehouse are constructed, a one-to-many training normal form is adapted, and the detection performance of anomaly detection in a training data scarcity scene is improved.

The invention provides a lightweight zero sample image classification method and system based on a state space model, and the method comprises the steps: inputting an unseen class test sample into a trained zero sample image classification network model, and outputting an unseen class image classification result; the zero sample image classification network model comprises a visual encoder, a semantic encoder, a multi-mode fusion module and a classification module. Wherein the visual encoder is used for inputting an image and outputting visual features; the semantic encoder is used for inputting a text and outputting semantic features; the multi-mode fusion module is used for inputting visual features and semantic features and outputting features fusing visual features and semantic features; and the classification module is used for inputting features fusing vision and semantics and outputting a classification result of the image. The method has the advantages that the overall classification precision of unseen categories can be kept at a high level, and the calculation complexity of the overall model can be effectively reduced.

The present invention provides a facial expression recognition method and system based on multi-cue associative learning, belonging to the technical field of computer vision. The recognition method comprises: inputting a pre-recognized facial image into a student model and / or a teacher model for facial expression recognition. A training method comprises: cropping a global facial sample image to obtain an upper half facial sample image and a lower half facial sample image; extracting cue features; acquiring adjacency matrices corresponding to the upper half facial sample image, the lower half facial sample image and the global facial sample image; fusing associated semantics by using a feature-level attention mechanism, so as to acquire the teacher model; supervising training of the teacher model by using a cross-entropy loss; and supervising training of the student model by using label distillation, KL divergence and a cross-entropy loss.

The present invention discloses a deep learning backdoor attack method and device based on an ordinal network and a medium, which belongs to the technical field of neural network security. The method comprises: obtaining a training sample image; generating an ordinal network based on the training sample image, wherein the ordinal network is used for representing an exact structure of an image; and generating a trigger by using the ordinal network. Through the above method, and the device and the medium for realizing the above method, the present invention uses the ordinal network to generate the trigger, improves the concealability of a poisoning sample compared with poisoning samples generated by other attack methods, and can promote the further research of a hidden backdoor attack defense method in the academic circle.