23 results about "Image reference" patented technology

This application provides a video image component prediction method and apparatus, and a computer storage medium. The method may include: acquiring a reference value set of a first image component of the current block; determining multiple first image component reference values ​​from the reference value set of the first image component; performing a first filtering process on the sample values ​​of the pixels corresponding to the multiple first image component reference values ​​respectively to obtain multiple filtered first image reference sample values; determining the image component reference value to be predicted corresponding to the multiple filtered first image reference sample values; determining the parameters of a component linear model based on the multiple filtered first image reference sample values ​​and the image component reference value to be predicted; performing a mapping process on the reconstructed value of the first image component of the current block according to the component linear model to obtain a mapping value; and determining the predicted value of the image component to be predicted of the current block based on the mapping value.

The present application relates to a task decomposition-based scene consistency image instruction generation method, and the generation method comprises the following steps: step S1, image analysis, object recognition, position and state recognition are performed; step S2, instruction understanding is performed, and set key information is extracted; step S3, feature extraction of a pre-scene image in a latent space is realized through a VAE encoder; step S4, a task decomposition generation step sequence is generated, and step-by-step description is performed; step S5, based on the step sequence, an instruction weight matrix with front and rear correlations is generated; the instruction weight matrix with front and rear correlations is designed based on the step sequence of step S4; step S6, based on the image reference of the pre-scene and the instruction weight, a visual instruction of the current scene is generated; the present application is reasonable in design, compact in structure and convenient to use.

This invention provides a method, apparatus, device, and storage medium for generating digital life forms from multiple reference images. The method includes: acquiring multiple target reference images corresponding to different attributes such as appearance, clothing, and scene; extracting features from each of the multiple target reference images to determine multiple target feature vectors; determining comprehensive constraints, which at least include a comprehensive condition vector and condition injection parameters; inputting the comprehensive constraints into a target generation model to generate an image, determining an initial generated image corresponding to the digital life form; and optimizing the quality of the initial generated image corresponding to the digital life form to determine the target generated image corresponding to the digital life form after quality optimization. The character generated by this invention is highly consistent with each reference image in appearance, clothing, and background, achieving a realism close to that of a photograph, and the implementation process is simple and low-cost.

Provided are a video image component prediction method and apparatus, and a computer storage medium. The method may comprise: acquiring a reference value set of a first image component of a current block; determining, from the reference value set of the first image component, multiple first image component reference values; respectively carrying out first filtering processing on a sample value of a pixel point corresponding to each of the multiple first image component reference values to obtain multiple filtered first image reference sample values; determining reference values, corresponding to the multiple filtered first image reference sample values, of an image component to be predicted; according to the multiple filtered first image reference sample values and the reference values of the image component to be predicted, determining a parameter of a component linear model; according to the component linear model, carrying out mapping processing on a reconstruction value of the first image component of the current block to obtain a mapping value; and according to the mapping value, determining a prediction value of the image component to be predicted of the current block. 20 24 28 72 36 27 D ec 2 02 4 A B S T R A C T 2 0 2 4 2 8 7 2 3 6 2 7 D e c 2 0 2 4

This application relates to a method, apparatus, computer device, and storage medium for remediating image vulnerabilities. The method obtains a hash fingerprint of the image layer marked as vulnerable, generated based on the image file change record and a timestamped salt value. Based on this hash fingerprint and the hierarchical dependencies encompassing the build relationships and version evolution relationships between image layers, candidate business images referencing the vulnerable layer are traced upwards, and related candidate newly built images and derived image layers are analyzed downwards. By comparing with currently running instances in the container cluster, the target image and image layer actually running are selected from the candidate objects. The target object is input into a preset remediation priority evaluation model to obtain a quantified remediation priority evaluation value, based on which a vulnerability remediation plan is determined. This method achieves accurate source tracing and dynamic convergence of the vulnerability's impact scope, and optimizes remediation decisions through priority evaluation, effectively improving the efficiency and targeting of vulnerability remediation.

The application relates to a minimum spanning tree-based SAR tomography reference net generation method and device. The method comprises the following steps: calculating amplitude deviation index values of all PS points in a remote sensing imaging area constituting an imaging reference net, selecting a PS point with the minimum amplitude deviation index value from all PS points as a starting reference point according to the amplitude deviation index values of all PS points, adding the starting reference point into an SPS point set, and initializing an SPS edge set as an empty set; the imaging reference net comprises a PS point set; according to a pre-set distance threshold, PS edges meeting the distance threshold are sequentially found in the SPS point set and the PS point set, and a candidate SPS edge set is set; the highest-quality edge is selected from the candidate SPS edge set and added into the SPS edge set, and the end points corresponding to the highest-quality edge are moved from the PS point set to the SPS point set, and finally, an SPS edge set and an SPS point set are output; and a SAR tomography reference net is constructed according to the final SPS edge set and the SPS point set. The method can generate a reference net which maximally covers a full scene.

This application provides an image classification method, apparatus, electronic device, and storage medium, belonging to the fields of artificial intelligence and digital healthcare technology. The method includes: acquiring and inputting training image features into an original image classification model including an original encoder and an original decoder; encoding the training image features using the original encoder to obtain training image vectors, the training image vectors including: image domain vectors, image semantic vectors, and image instance vectors; decoding the training image vectors using the original decoder to obtain image prediction categories; decoupling the training image vectors into a decoupling model to obtain image decoupling information; adjusting the parameters of the original image classification model according to preset image reference categories, image prediction categories, and image decoupling information to obtain a target image classification model; and inputting target image features into the target image classification model to perform image classification to obtain the target image category. This application can improve the accuracy of image classification.

This invention discloses a two-stage large-scale model report generation method and apparatus that integrates multimodal data. The method includes: reading a structured text knowledge set, a quantized data subset, and a semantically named image set; generating a composite dictionary and serializing it into a context data string; inputting the context data string and a predefined report template string into a first language model to generate initial report text; generating image-encoded data based on the image set; inputting the initial report text and image-encoded data into a second visual language model to generate a revised report text; verifying and aligning image reference paths in the revised report text; and outputting the final report text. This invention achieves fact-checking and positive feature enhancement through two-stage large-scale model collaboration, utilizes structured template constraints for output, and automatically segments the data into multiple files, solving the problems of insufficient use of visual evidence, poor factual accuracy, and limited output format in existing technologies.

The application discloses an image processing method, system, chip and storage medium. Firstly, the application reads local input image data, corresponding local image reference data and local noise characteristic data from an on-chip cache, then performs fusion and noise reduction processing according to the three kinds of local data to obtain local noise-reduced input image data, local fusion data and local updated noise characteristic data, merges the local noise-reduced input image data to obtain noise-reduced input image data, and updates the on-chip cache by merging the local fusion data and the local updated noise characteristic data. Thus, the application implements an image noise reduction technical solution adapting to hardware constraints under an ISP architecture mainly using stream processing, without relying on external DDR storage, with limited on-chip resources, which can significantly reduce system bandwidth occupation and power consumption, and can also realize light and effective noise suppression while ensuring real-time performance.

This invention provides a method for remote sensing image referential segmentation, relating to the field of artificial intelligence technology. The method includes: generating counterfactual text with semantics opposite to the original referential text corresponding to the remote sensing image; aligning the original referential text with the visual features of the remote sensing image to obtain basic fusion features; simultaneously aligning the counterfactual text with the visual features to obtain counterfactual fusion features; determining the difference features between the basic fusion features and the counterfactual fusion features, and fusing the difference features with the basic fusion features to obtain enhanced alignment features; and decoding the enhanced alignment features to obtain the referential segmentation result of the remote sensing image. This invention, by generating counterfactual text with semantics opposite to the original referential text and performing dual-path feature alignment and difference fusion, effectively eliminates the influence of confounding variables at the visual and linguistic levels, allowing segmentation to focus on causal features truly relevant to the original referential text, significantly improving segmentation accuracy and interpretability in complex remote sensing scenarios.

The invention discloses an image processing method and device, electronic equipment, a storage medium and a program product, and relates to the technical field of image processing. According to the image processing method, when water ripples exist in a short-exposure sub-frame, the long-exposure sub-frame can be used as a standard sub-frame for short-exposure sub-frame imaging reference due to the fact that the difference of exposure energy among multiple rows of pixels in a long-exposure sub-frame adjacent to the short-exposure sub-frame is small; therefore, the exposure compensation value for performing exposure compensation on the short exposure sub-frame can be determined based on the difference between the exposure quantity and the brightness characteristic between the short exposure sub-frame and the standard sub-frame, so that water ripples in the short exposure sub-frame are removed by performing exposure compensation on the short exposure sub-frame. Based on the short exposure sub-frame after exposure compensation, the imaging quality of the output image generated by image fusion of the electronic equipment is improved, and the user experience is further improved.

Image generation device for a vehicle, with: an image reference unit (F1) configured to receive multiple camera images obtained from multiple cameras configured to photograph the vehicle's surroundings; an image synthesis unit (F7) configured to project data from the majority of camera images onto a virtual projection surface corresponding to the vehicle's environment and to generate a synthetic image showing the vehicle's environment as viewed from a virtual viewpoint, using the data projected onto the projection surface; and a driving environment determination unit (F62) configured to determine, based on a signal from another on-board device of the vehicle, whether a driving environment of the vehicle is terrain or on a road, wherein the image synthesis unit (F7) is configured to change the shape of the projection surface used to generate the synthetic image, depending on whether the driving environment determination unit (F62) determines that the driving environment is terrain, the image synthesis unit (F7) is configured to generate the synthetic image by using a road projection surface with a predefined shape as the projection surface when the driving environment determination unit (F62) determines that the driving environment is on a road, and The image synthesis unit (F7) is configured to produce the synthetic image by using a terrain projection surface with a shape different from the shape of the road projection surface when the driving environment determination unit (F62) determines that the driving environment is terrain.

This application provides a video image component prediction method and apparatus, and a computer storage medium. The method may include: acquiring a reference value set of a first image component of the current block; determining multiple first image component reference values ​​from the reference value set of the first image component; performing a first filtering process on the sample values ​​of the pixels corresponding to the multiple first image component reference values ​​respectively to obtain multiple filtered first image reference sample values; determining the image component reference value to be predicted corresponding to the multiple filtered first image reference sample values; determining the parameters of a component linear model based on the multiple filtered first image reference sample values ​​and the image component reference value to be predicted; performing a mapping process on the reconstructed value of the first image component of the current block according to the component linear model to obtain a mapping value; and determining the predicted value of the image component to be predicted of the current block based on the mapping value.

The present application provides a kind of lightweight remote sensing image reference segmentation method, system, equipment and medium, belong to computer vision and remote sensing intelligent processing field, method includes: Taylor pruning is carried out to visual encoder, and the multiscale visual features of remote sensing image are extracted using the visual encoder after pruning;Text encoder is used to extract the reference text features of natural language description;Through point-by-point weighted attention mechanism, the visual features are multiplied with the reference text features element by element, to obtain multiscale cross-modal fusion features;The cross-modal fusion features are fused in a top-down and bottom-up manner using a bidirectional path enhancement module, to obtain enhanced fusion features;Lightweight feature pyramid decoding head is used to decode and segment prediction, to obtain reference segmentation result.The present application effectively reduces the model parameter quantity and the calculation cost, improves the cross-modal alignment and multiscale feature fusion capability, and balances the lightweight and segmentation accuracy.

Provided is a medical image photographing support device capable of determining whether or not an installation state is eccentric, a medical image photographing support device, a method of operating a medical image photographing support device, a recording medium, a program product, and a medical image photographing system. The medical image photographing support device performs the following processes: acquiring a first photographing image of a subject obtained by photographing the subject placed on a table using a photographing device; detecting an adjustment reference provided on the table from the first photographing image; deriving a difference between a photographing image reference provided in the first photographing image and the adjustment reference; and determining eccentricity of an installation state of the photographing device based on the difference.

This invention provides a topology-guided and visual feature embedding image editing method, apparatus, and device, relating to the field of image processing technology, to address the defect of poor quality of edited images generated during image editing. The method includes: acquiring the image to be edited and editing intent information; selecting a target editing strategy from a set of preset editing strategies based on the editing intent information, the multiple editing strategies including at least a topology-guided editing strategy and an image reference editing strategy; invoking a diffusion model corresponding to the target editing strategy, and generating an edited target image based on the image to be edited and the editing intent information; fine-tuning a base model based on multiple samples containing the target object to obtain a diffusion model corresponding to the topology-guided editing strategy; and expanding the input channels of the base model to obtain a diffusion model corresponding to the image reference editing strategy, the expanded channels being used to input a reference image including the target object.

The application discloses a kind of based on the dynamic joining method and system of the reference point of PCB board of DLL, method S1, the ID attribute value of reference point is set;S2, read the CAD file of PCB board, parse CAD element;S3, write graphical reference point calculation program;S4, write image reference point calculation program;S5, write the use program of reference point;S6, generate DLL dynamic link library file;S7, establish the mapping of ID and DLL, and different reference point generation method is selected by ID.The application overcomes the waste of traditional multiple modification program code mode to equipment and human resources etc., using the idea of generating DLL, greatly reduces the programming workload of reference point, can conveniently and quickly realize the dynamic joining of reference point, improves the test efficiency of reference point.

This invention relates to a method for automatic document format detection, belonging to the field of office automation technology. The invention selects the document information to be detected and directly reads relevant document information using a multi-threaded serial approach; it also uses a multi-threaded serial approach to obtain specific document information, including page layout information and paragraph details; and it uses a multi-threaded parallel approach to detect formatting issues in titles, images, figure captions, image references, tables, table headers, cells, table titles, table references, and body text. The detection results are categorized and displayed on the interface, exported to Excel, added to the document as annotations, and element styles corrected. Manual corrections are made to the detected issues, or the corrected document is confirmed. This invention offers a lightweight, user-friendly, flexible, comprehensive, secure, and fast retrieval process with accurate detection results.

This invention discloses a remote sensing image referencing segmentation method and system, addressing the technical problems of insufficient long-range dependency modeling capability and low multi-scale target perception accuracy in the extraction of complex features from remote sensing images. The method includes: encoding the input image using a state-space model-based image encoder to obtain multi-scale visual features; encoding the input text using a text encoder to obtain a text feature sequence and a global sentence vector; expanding the spatial dimension of the global sentence vector to obtain global text semantic features; generating local interaction mapping features based on the multi-scale visual features and the global sentence vector to obtain multi-modal hybrid features; enhancing the multi-modal hybrid features using multi-scale dual-dimensional scanning to obtain a fused enhanced feature map; decoding the fused enhanced feature map using a state-space model-based decoder to obtain the final decoded output features; and obtaining the segmentation result based on the final decoded output features.

The present application relates to the field of image data processing, and provides a steam boiler gas leakage detection method based on image processing, which comprises the following steps: setting an image acquisition unit to jointly cover a predetermined leakage risk area, acquiring corresponding reference images under different production conditions, and establishing an image reference library covering a target area; dividing a boiler operation time axis into multiple production states, and calling corresponding reference images for different states, performing difference, partition comparison and motion detection on real-time acquisition images, and obtaining suspected leakage cloud area under corresponding states; arranging a condensate water indicating patch, acquiring patch images, extracting color, brightness and texture features to form a patch humidity change curve, performing multi-source information fusion on time sequence features and production time states, and outputting corresponding steam leakage risk levels; generating alarm information according to the steam leakage risk levels, and recording and archiving in a monitoring and management system for energy consumption analysis and risk tracing.

Provided are a medical image capturing support apparatus, an operation method of a medical image capturing support apparatus, a program, and a medical image capturing system that can determine whether or not there is misalignment of a mounting state of an imaging apparatus that acquires a captured image of a subject. A medical image capturing support apparatus acquires a first captured image of a subject placed on a bed, the first captured image being acquired by imaging the subject using an imaging apparatus, detects an adjustment reference defined on the bed from the first captured image, derives a difference between a captured image reference set for the first captured image and the adjustment reference, and determines misalignment of a mounting state of the imaging apparatus based on the difference.

This invention discloses a method, apparatus, electronic device, and storage medium for generating responses to power transmission operation queries, belonging to the field of artificial intelligence technology. The method includes: acquiring a set of supporting documents for power transmission operation to be processed and a query question; preprocessing the document set to extract an initial set of text paragraphs, an initial set of images, image reference identifiers, and image access addresses; segmenting the text using image reference identifiers as positioning points to generate a pre-description paragraph and a post-analysis paragraph, and constructing a knowledge feature vector based on the image reference identifiers; constructing a question feature vector based on the query question; selecting target knowledge feature vectors and determining a target context set; inputting the query question and the target context set into a text generation model to generate initial response text, and combining the image access addresses to output the final response result. By implementing this invention, the problem of inaccurate response generation for power transmission operation queries in existing technologies can be solved.

This invention belongs to the field of artificial intelligence and provides a method, apparatus, device, and storage medium for document review based on OCR and NLP. The method includes: acquiring images of underwriting documents and image reference information; identifying face regions and character regions from the document images; when a document image is determined to be valid based on the proportion of the face region, identifying target characters and content characters using a preset OCR model, and determining a target dictionary based on the title characters of the target characters; inputting the target dictionary and content characters into a preset NLP model to identify the document review result. According to the technical solution of this embodiment, the validity of document images can be initially screened based on face proportion, eliminating the need for manual verification and reducing labor costs. Furthermore, after determining that the document image is valid, the OCR model and NLP model analyze the document review result, realizing automatic document image review and improving work efficiency.