641 results about "Image database" patented technology

The invention belongs to the technical field of information, and relates to a semantic guidance pedestrian re-identification method and system based on text prompt. The method comprises the following steps: inputting a training image into a visual encoder to obtain visual embedding; mapping the visual embedding into a text space by using a reverse network to obtain a pseudo token, and integrating the pseudo token into a natural language sentence to obtain a language prompt for an input image; inputting the language prompt into a text encoder to obtain text embedding; training a multi-modal interaction module by utilizing visual embedding and text embedding; and inputting a query picture into the trained multi-modal interaction module to obtain a feature vector fusing vision and text information, and executing similarity retrieval in a pedestrian image database by using the feature vector fusing vision and text information to obtain a pedestrian re-identification result. According to the method, the retrieval performance on an existing pedestrian re-identification data set is remarkably improved.

Provided is a system and method for digital content generation and recommendation. A first plurality of stories are collected from various content sources over a network. A second plurality of stories are selected from the first plurality of stories such that a degree of similarity between any two stories of the second plurality of stories is below a determined threshold. A summary and a headline are generated for a first story of the second plurality of stories. A set of images are selected, from an image database, for the first story based on the generated summary and headline. A multimedia item including the summary, the headline, and a first image of the set of images is created. The multimedia item is presented on a lock screen interface of a user device, displaying the generated summary, the generated headline, and the first image on the lock screen interface.

The invention discloses an industrial visual detection system based on multi-modal fusion, which belongs to the technical field of industrial data acquisition, processing and processing, and comprises a multi-modal data acquisition and calibration module, a scene image capture module, a visual information synthesis module, an algorithm optimization and training module and a multi-modal image output module, the multiple modules are integrated into the industrial visual detection system for multi-modal data acquisition, processing and fusion, high-quality original data are provided for subsequent image analysis, images of industrial scenes can be automatically captured under different illumination conditions, an original image database containing various illumination and scene conditions is constructed, and the visual detection system has the advantages of being simple in structure, convenient to operate and high in practicability. The system can operate stably, high-precision analysis and processing of the image data set in the industrial scene are realized, the corresponding fusion image is generated, the industrial visual inspection report is generated based on the fusion image, and the adaptability and reliability of the industrial visual inspection system are improved.

The invention discloses a medical image report generation method based on multiple agents. The method comprises the following steps: step 1, constructing an image segmentation agent, a cue word retrieval agent, a report generation agent and an image database; 2, inputting the medical image into an image segmentation agent, and generating a focus label through the image segmentation agent; 3, inputting the focus label into a cue word retrieval intelligent agent, and obtaining an example report related to the focus label by the cue word retrieval intelligent agent; and 4, taking the focus label and the example report as input of a report generation agent, outputting a preliminary image report by the report generation agent, and then introducing a terminology condition constraint and reinforcement mechanism to correct the preliminary image report to generate a final image report. According to the invention, the workload of doctors can be reduced while the image analysis report which is accurate and meets the medical standard is generated.

The invention relates to the technical field of remote medical treatment, in particular to a remote ultrasonic diagnosis support system based on cloud computing, which comprises an echo feature analysis module, a dynamic focus screening module, a case image comparison module, a focus trend classification module and a remote diagnosis support module. According to the method, by analyzing echo signals in a remote ultrasonic image data stream and calculating the echo change rate between image frames, the dynamic characteristic capturing capacity of a lesion area is improved, calculation of a gray gradient change value is combined, a suspected lesion area conforming to a change rate stable interval is screened, lesion shape edge curve contour parameters are utilized, and the dynamic characteristic capturing capacity of the lesion area is improved. The method comprises the following steps: performing matching analysis with an ultrasonic case image database, enhancing the accuracy of lesion category attribution, analyzing the morphological change trend of a lesion area, combining trend stability index calculation, realizing lesion evolution classification, ensuring the reasonability of lesion type division, and marking a progressive lesion area, so that remote diagnosis is more intuitive; and the auxiliary decision-making capability is improved.

The invention provides a food freshness evaluation method and system based on image processing. The method comprises the following steps: acquiring a multi-view image of food, transmitting the image to a central processing center, setting personalized image acquisition parameters, and forming a food image database; based on the food image database, performing high-precision dynamic monitoring on color and texture changes of the food in different time periods, and generating a preliminary evaluation result; according to the preliminary evaluation result, combining food types, storage conditions and historical sales data, performing prediction modeling on the freshness of the food to obtain a final freshness evaluation report; and utilizing the freshness evaluation report to adjust an inventory strategy in real time, optimizing a food distribution plan, and generating an optimized management process. According to the technical scheme provided by the invention, the accuracy of food freshness evaluation is improved.

The invention discloses a brain tumor area automatic segmentation method and system based on deep learning, and belongs to the technical field of image segmentation, and the method comprises the following steps: obtaining a disclosed brain tumor segmentation data set, and constructing a multi-modal brain tumor medical image database; performing data preprocessing on the multi-mode brain tumor medical image database; dividing the preprocessed multi-modal brain tumor medical image database into a training set, a verification set and a test set in proportion; a deep learning model LG-FPN-TransBTS based on the improved TransBTS is designed, and the deep learning model LG-FPN-TransBTS is designed; and training, evaluating and applying the model. The system comprises a data preprocessing module, a data set division and secondary processing module, a deep learning model construction module, a model training and optimization module and a model evaluation and application module. According to the method, the precision and robustness of brain tumor image segmentation are remarkably improved.

The invention discloses an airborne infrared-visible double light combined building outer wall hollowing identification and positioning method. The method comprises the following steps: designing flight parameters of an unmanned aerial vehicle; designing a calibration checkerboard; shooting an infrared thermal image and a visible light image of the checkerboard calibration target, recording the images as dual-light images for calibration, and calibrating and calculating internal and external parameters and a distortion coefficient of a camera; collecting images of all external facades of the target building; carrying out distortion removal processing on the dual-light image to form a dual-light image database for detection; the registration between the dual-light images synchronously shot by the unmanned aerial vehicle is realized; realizing registration between the building facade view and the visible light image; determining a correct hollowing area identification result, and marking the infrared thermal image corresponding to the correct hollowing area identification result as a hollowing infrared thermal image; calculating a pixel coordinate range of the hollowing area; and accurate positioning of the hollowing area is realized. According to the method, the visible light image serves as a medium, registration of the infrared thermal image and the macroscopic 3D model is achieved, the problem that calibration and registration are difficult due to the fact that the imaging quality of the infrared thermal image is low is solved, and high readability and user friendliness of an output result are guaranteed.

The invention provides an infrared and visible light image end-to-end registration method based on phase consistency enhancement, and belongs to the technical field of electric digital data processing. The invention discloses an infrared and visible light image end-to-end registration method based on phase consistency enhancement, which comprises the following steps of: firstly, preprocessing infrared and visible light source images in an image database, and dividing the infrared and visible light source images into a training set and a test set; and constructing a feature extractor based on a phase consistency attention mask, realizing cross-modal consistency feature extraction, and focusing an important space region by using phase information. A self-defined ResSobeNet network is adopted for parameter estimation, and Haar wavelet transform is used for replacing pooling operation, so that information loss in a traditional down-sampling method is avoided. A double-branch generator structure is designed, affine parameters and an optical flow field are generated respectively, meanwhile, constraint on rigid registration and non-rigid registration is achieved, and finally a registration result is obtained through resampling of a space conversion module. According to the method, the problems that cross-modal consistent features are difficult to extract in the infrared and visible light image registration process and explicit registration steps are complicated and difficult to generalize are solved, and the method has good universality and practicability.

The invention relates to the technical field of substation defect detection, in particular to a substation defect detection method and system and a storage medium. The method comprises the following steps: acquiring a substation equipment image by adopting real-time image acquisition equipment, and establishing an equipment image database; performing denoising and edge detection through image processing software to generate preprocessed image data; detecting an abnormal area of the substation equipment through feature extraction and texture analysis, and training a defect recognition model; after real-time image information is input into the model, an abnormal defect area of the transformer substation is recognized, and fault risk gradient deterioration is estimated; calculating a risk index based on a pre-estimation result, and if the risk index exceeds a set threshold value, triggering an early warning function, providing defect types and positions, and suggesting repair measures; according to the invention, the transformer substation defect detection method is optimized, so that the transformer substation defect identification is more accurate.

The present disclosure describes approaches to camera re-localization using a graph neural network (GNN). A re-localization model includes encoding an input image into a feature map. The model retrieves reference images from an image database of a previously scanned environment based on the feature map of the image. The model builds a graph based on the image and the reference images, wherein nodes represent the image and the reference images, and edges are defined between the nodes. The model may iteratively refine the graph through auto-aggressive edge-updating and message passing between nodes. With the graph built, the model predicts a pose of the image based on the edges of the graph. The pose may be a relative pose in relation to the reference images, or an absolute pose.

The invention discloses an accurate feeding system and method based on machine vision. The system comprises an image acquisition module, an image processing module, a satiation degree evaluation module, an automatic feeding machine and a control module. The image acquisition module is used for acquiring a water surface spray image of the fishpond in real time; the image processing module is in signal connection with the image acquisition module and is used for carrying out segmentation and feature extraction on the water surface spray image acquired in real time by adopting a convolutional neural network to obtain spray features of the water surface spray image, and the spray features comprise the size and the duration of the spray; the water spray features before feeding, during feeding and after feeding are stored in a feeding water spray image database; and the satiation degree evaluation module is used for evaluating the ingestion active degree and the satiation degree of the fish school according to the ingestion spray characteristics on the basis of a satiation degree evaluation model. According to the method, the feeding speed is adjusted according to the collected and analyzed water spray characteristics on the water surface, so that the feeding speed is coordinated with the feeding speed, precise feeding is achieved, and residual feed is reduced.

The invention discloses an adenoid body recognition method and device based on image instance segmentation, and the method comprises the steps: selecting a skull side position film image in a medical image database, and carrying out the marking of the skull side position film image to clearly know the position of a key structure; preprocessing each image to generate a training sample; and training an instance segmentation model by using the samples, wherein the model comprises a backbone network, a feature fusion module and a detection head. The backbone network is responsible for preliminary feature extraction, and the C2f module further performs deep processing on feature maps, generates multi-scale feature maps and fuses the multi-scale feature maps into a comprehensive feature map. And the feature fusion module adopts a path aggregation network to integrate all scale feature maps. And the detection head receives the fusion feature map, generates a candidate frame and performs classification prediction. And adjusting the candidate frame according to the prediction result to ensure accurate identification of the adenoid body, and outputting the adjusted candidate frame and the category label as a final identification result. The problem of how to improve the accuracy and efficiency of diagnosis by accurately segmenting and identifying adenoid bodies and related structures thereof is solved.

The invention relates to the field of intelligent city planning, and discloses a street space quality intelligent evaluation and image generation method and system based on deep learning, and the method specifically comprises the following steps: movably collecting a street space image, and recording the geographic position information; performing perspective correction and multi-view image segmentation on the acquired street image data, and performing spatial matching with geographical location information to obtain a fused image database; and performing semantic segmentation on the image in the fused image database based on a deep learning model, calculating objective and subjective visual perception indexes and facility function indexes of the image, and generating a street space quality evaluation result. The method solves the problems that in the prior art, evaluation dimensions are limited, and an automatic response generation mechanism is lacked, and has the advantages of being intelligent, efficient and comprehensive in subjective and objective evaluation.

The invention relates to the technical field of computer vision and image processing, and provides a depth synergy graph Hash image retrieval method based on attention enhancement, which comprises the following steps: acquiring a to-be-retrieved image; extracting image features of the to-be-retrieved picture; performing weighting processing on the image features by using an attention mechanism to obtain weighted image features; encoding the weighted image features into a hash code; hash codes of pre-stored images in an image database are obtained, Hamming distances between the Hash codes of the to-be-retrieved images and the Hash codes of the pre-stored images are calculated, the pre-stored images are sorted according to the sequence of the Hamming distances between the Hash codes of the to-be-retrieved images and the Hash codes of the pre-stored images from small to large, and the first multiple pre-stored images are output to serve as similar images of the to-be-retrieved images; the Hash codes obtained through the image retrieval method are high in discrimination capability and high in image retrieval precision.

The invention discloses an underwater concrete structure crack detection method based on edge calculation and deep learning, and the method comprises the steps: carrying out the pixel preprocessing of underwater concrete structure crack image data in an image database, and obtaining a preprocessed image; constructing a lightweight image segmentation network to perform image enhancement processing on the preprocessed image to obtain an image enhancement data set; constructing an ASF-ODConv-YOLOv8 network model, and performing network training and verification according to the image enhancement data set to obtain an intelligent crack recognition algorithm model; based on the crack intelligent identification algorithm model, obtaining a predicted crack category label of the underwater concrete structure crack according to the test set; and constructing a digital twinborn model of the underwater concrete structure crack so as to detect the underwater concrete structure crack. The problems that a traditional recognition method has limitations in the aspects of space-time coverage, environmental adaptability, recognition efficiency and precision, depends on subjective judgment and is prone to causing missing recognition and wrong recognition are solved.

To make a data-base of images associated with a natural descriptive text such that it matches a retrieval sentence entered by a person.SOLUTION: A descriptive text generating part 12 of a vehicle image analyzer 1 identifies a scene captured in an image received from a camera, reads recognition necessity information on a target in the image for each of the identified scene from a necessity table 14A, identifies the target being specified as necessary to be recognized in the read recognition necessity information, generates a descriptive text for each target from the recognition result, generates a situation descriptive text of the image on the basis of the identified scene and the descriptive text for each target, and memories the image associated with the situation descriptive text of the image into a descriptive text DB 15.SELECTED DRAWING: Figure 1

The invention discloses a big data hypergraph network modeling method and device, computer equipment and a product, aims to solve the problem that efficiency and accuracy are difficult to consider when large-scale data of intellectual property is processed in the prior art, and aims at the characteristics of high increment, high timeliness and multiple modes of the intellectual property data. The method comprises the following steps: analyzing and concluding passed intellectual property files, classifying according to corresponding professions, and establishing a corresponding intellectual property text library; constructing a text corpus based on intellectual property by using a BERT algorithm model, performing capsule graph neural network processing on the image information of the corresponding category, obtaining a correlation between image objects and text attributes thereof, and constructing an image database; capturing complex correlation characteristics in the text by utilizing the characteristics of the graph convolutional neural network; and constructing a hypergraph neural network model. The invention provides a big data hypergraph network modeling method capable of processing high-increment, high-timeliness and multi-modal data.

The invention discloses a buried hill core porosity calculation method based on a high-precision repair CT image, and the method comprises the steps: S1, obtaining a micron CT three-dimensional gray image of a buried hill fracture reservoir plunger sample in a research region, and obtaining a three-dimensional digital core model of the buried hill fracture reservoir plunger sample based on the micron CT three-dimensional gray image; s2, obtaining nanometer CT three-dimensional grayscale images of the sub-samples at different positions; s3, establishing a buried hill fractured reservoir core image database under different resolutions, learning and training the image through a residual network model to obtain a high-precision repaired micron CT image, and simulating and calculating the porosity through threshold segmentation; and S4, determining the actually measured porosity of the buried hill crack sample through an indoor experiment, and comparing the actually measured porosity with the simulated calculation porosity. According to the method, the pixel boundary contact information of the micron CT image is recovered through the deep residual network model, the restoration quality of the image is improved, and the simulated porosity is closer to an actual value.

The invention provides an MRI image enhancement method and system based on multi-modal fusion, and belongs to the technical field of image enhancement, and the method comprises the steps: building a medical image database, and obtaining a diagnosis text; original MRI image data are obtained, and structured processing is carried out on T1W / T2W and ADC image data; a fusion enhancement algorithm model pool is established, a forward enhancement model, a reverse optimization model and a CT auxiliary enhancement model are included in the fusion enhancement algorithm model pool, and matching operation is carried out; if the matching is successful, starting a CT auxiliary enhancement model, and performing fusion enhancement with the forward enhancement model; if the matching fails, independently starting a forward enhancement model to carry out fusion enhancement; forward outputting the enhanced MRI image data into the medical image database, reversely outputting the enhanced MRI image data into the reverse optimization model, and regularly updating and optimizing the forward enhancement model and the CT auxiliary enhancement model; a method for enhancing an MRI image based on CT image selectivity is provided.

The present disclosure includes a method for generating synthetic on-person screening images including providing a database of labeled on-person screening images, wherein a labeling of a respective on-person screening image in the database indicates: the presence of an object in the image, and, in case the object is present, at least one of: a classification, a location, a size, a material and a weight of the object. The method includes training a generative AI model by feeding at least some of the labeled on-person screening images in the database to the generative AI model; receiving an input where the input includes anatomic information of a person, information on a presence or absence of at least one further object, and / or information on a location of the further object relative to the person; and generating at least one synthetic on-person screening image with the generative AI model based on the input.

The invention discloses an airspace radiation normalization method, system and device for large-area multi-source data. The method comprises the following steps: generating a multi-cause radiation normalization image database based on a remote sensing observation image set and forming sample image pairs; constructing a generative adversarial network, carrying out fusion processing on the sample image pair through a super-resolution module to form a fusion definition image, forming a radiation consistent image through a style migration module, optimizing the radiation consistent image through a normalization module to form a normalized image to complete optimization, and obtaining a normalized adversarial pre-training model; training the normalized confrontation pre-training model through a multi-cause radiation normalized database to obtain a normalized confrontation model; and processing the spatial radiation to-be-normalized image based on the normalized adversarial model to obtain a corresponding spatial radiation normalized image. According to the method, the incidence relation of the data space-time spectrum can be deeply mined, and the problems of inconsistency of complex nonlinear radiation and global overall optimization are solved.

The invention discloses a pre-splitting blasting half-wall porosity intelligent identification system and method, and the system comprises a front-end display platform and a rear-end operation system which carry out data interaction through a computer, and the operation result information is displayed and checked through the front-end display platform; the front-end display platform consists of a menu bar, a toolbar, a display window and a command line; the rear-end operation system mainly comprises a half-wall hole image database, an image denoising module, an image segmentation module, a data statistical analysis module, an identification calculation module and the like, half-wall hole photo images are collected on steps on the two sides of the platform through the unmanned aerial vehicle oblique photography technology with or without the RTK function, a three-dimensional image database marked with half-wall hole information is established, and the three-dimensional image database is used for storing the half-wall hole information. An intelligent half-wall porosity recognition algorithm based on computer deep learning is developed, accurate recognition of the half-wall porosity is achieved, the traditional presplitting blasting quality evaluation efficiency can be greatly improved, and the method can be widely applied to slope presplitting blasting effect evaluation in the fields of surface mines, water conservancy and hydropower, transportation and the like.

The invention provides an image searching method, which comprises the following steps of: when a new label is added to an image in a first image database, acquiring a first semantic feature of the new label; based on the first semantic feature, performing image retrieval in a first image database to obtain a first retrieval result and a confidence coefficient of the first retrieval result; if the confidence coefficient is smaller than the preset confidence coefficient, performing secondary retrieval in the first retrieval result based on the existing tag in the first retrieval result to obtain a second retrieval result; based on the second retrieval result, adding the newly added tag into an image corresponding to the secondary retrieval result to obtain a second image database; and after an image search instruction of the user is obtained, performing image retrieval in the second image database based on the image search instruction. The problems that in the database maintenance process, a new label needs to be added, the workload of adding the label again to the data is increased while the new label is added, and the database is difficult to maintain in an existing method are solved.

Camera positions can be recovered for video frames. Keypoint tracks can be generated for a first video frame and subsequent video frames. Metric depth can be determined for the video frames. An image databased can be generated using the keypoint tracks and the metric depth. The image database can include the video frames. A first video frame having a highest number of keypoint tracks can be identified. A first projection into a three-dimensional model can be generated for the first video frame and a first matching video frame. A camera position can be generated using the first projection. A second video frame having a next-highest number of keypoint tracks can be identified. A second projection into a three-dimensional model can be generated for the second video frame and a second matching video frame. A camera position can be generated using the second projection.

The invention discloses a wind turbine blade fault diagnosis method based on aerodynamic noise and image recognition, and relates to the technical field of wind turbine blade fault diagnosis, and the method comprises the steps: building a plurality of typical fault models of a wind turbine blade based on a numerical simulation platform, and extracting simulation acoustic image data; the position of the aerodynamic noise source is positioned through an acoustic camera, and corresponding blade image information is collected by an unmanned aerial vehicle; and carrying out feature training on the image sample and the acoustic image data by using a deep neural network and executing fault matching judgment. According to the method, a fault acoustic image database under multiple working conditions is constructed through numerical simulation, outfield data acquisition is performed in combination with an acoustic camera and an unmanned aerial vehicle, and joint feature training and fault matching of acoustic frequency spectrums and images are realized by using a deep neural network. The method can break through the limitations of the prior art in the aspects of real-time performance, accuracy and intelligence, and achieves the efficient and reliable diagnosis of the fault of the wind turbine blade.

The invention provides a mask prior-based two-stage infrared and visible light image fusion method, belongs to the technical field of infrared and visible light image fusion, and solves the problem that scene information cannot be comprehensively expressed by an infrared or visible light single image. The method comprises the following steps: 1, preprocessing infrared and visible light source images in an image database, and dividing the infrared and visible light source images into a training set and a test set; 2, training a salient target detection network to obtain a trained salient target detection network model; and 3, inputting the infrared source image into the salient target detection network to obtain a mask image. And 4, constructing an image fusion network, and training the image fusion network to obtain a trained image fusion network model. And 5, testing the images in the test set by using the image fusion network model to obtain a fused image. According to the invention, based on the mask priori knowledge, the saliency information in the infrared image can be effectively extracted, so that the obtained fusion image can better highlight the target information, and the influence of environments such as strong light on the fusion image is weakened.

The invention discloses a battlefield semantic situation rapid generation method based on image analysis, and belongs to the technical field of battlefield situation generation. The method comprises the following steps: dividing a battlefield area into a plurality of local battlefields; reconnaissance is carried out on each local battlefield area; generating a multi-source heterogeneous image database; performing time-space relationship unification on the multi-source heterogeneous image database; constructing a battlefield situation element composition system; constructing a battlefield situation element extraction technology system; carrying out adaptive extraction on battlefield situation elements; generating a local battlefield situation; a global battlefield situation is generated based on the local battlefield situation. The invention provides a battlefield situation element extraction framework based on deep learning, designs a battlefield situation generation method based on deep learning, and provides a conflict situation fusion method and a battlefield region division method based on an attention mechanism and a complete space-time relationship unification mode. A whole set of battlefield semantic situation adaptive generation method is constructed, and decision-making assistance can be provided for direct commanders and fighters.

The invention discloses an image secret type cloud storage method and device based on feature block encryption, and the method comprises the steps: a cloud server receives and stores an encrypted image, an encrypted image feature set and a feature inner product, and forms an image database; the cloud server receives a feature ciphertext query feature uploaded by the authorized user when the authorized user needs to query the image; the cloud server performs retrieval according to the received ciphertext query features, separately calculates the similarity between color and texture sub-features in the ciphertext query features and features in each encrypted image feature set in the image database by using a hierarchical earth moving distance algorithm, and calculates the comprehensive similarity corresponding to each encrypted image; and screening out a plurality of encrypted images according to the comprehensive similarity and returning the encrypted images to the authorized user. According to the method, safe and efficient search of the image can be realized, and the image retrieval performance is improved.

The invention discloses a pathological image self-supervised dyeing normalization method and device, and belongs to the technical field of combination of digital pathology and image processing, and the method comprises the steps: constructing a pathological image database; a dyeing normalization model is designed, the dyeing normalization model comprises a public feature domain extraction module, a content perceptron, a scene perceptron and a fusion module, and the dyeing normalization model can perform color conversion on the pathological image to obtain a dyed normalized image; a self-supervised loss function is constructed based on the pixel level difference of the input original pathological image and the dyed and normalized image, and a dyeing and normalization model is trained. The model can perform adaptive dyeing style normalization on any image of an unknown pathological data center towards a selected pathological data center image.