19 results about "Image fusion" patented technology

This invention relates to the field of medical image processing technology, specifically to a deep learning-based preoperative multimodal image fusion and evaluation system for stroke, comprising: acquiring multimodal image sequences; determining the vascular collapse index at the center point of each blood vessel based on the edge features of the DSA angiography sequence; determining the mean cellular edema index and vascular collapse index of each local region based on MRI images, thereby determining the degree of cytotoxic edema in each local region; mapping the local region of the MRI images onto CT images, and determining the edge dissolution risk coefficient of each local region based on the edge features of each local region in the CT image sequence, combined with the degree of cytotoxic edema; using the multimodal image sequences as input, dynamically adjusting the edge intensity map using the edge dissolution risk coefficient to obtain the multimodal preoperative lesion region fusion result to assist doctors in evaluation. This invention effectively ensures the authenticity of the three-dimensional fusion result.

The application provides a tumor radiotherapy multi-modal image fusion method and system, which extracts an initial target region contour of a tumor and a set of all high metabolic voxels; determines the shortest spatial distance from each voxel in the set of high metabolic voxels to all contour points on the initial target region contour, and then generates a distance statistical histogram; identifies a far-end distance point cluster that is significantly separated from a main distribution cluster, separates a corresponding spatial outlier voxel cluster from the set of high metabolic voxels; on the initial target region contour, determines continuous contour points closest to the spatial outlier voxel cluster as a to-be-corrected contour segment; determines a displacement vector of each point on the to-be-corrected contour segment to the spatial outlier voxel cluster, adjusts the spatial position of the to-be-corrected contour segment according to the displacement vector, and obtains an optimized target region contour of the tumor. The scheme of the application can avoid definition distortion of the target region contour caused by main noise interference.

This invention relates to the fields of remote sensing image processing and agricultural information technology, and in particular to a method for refined extraction of winter wheat planting areas based on Gaofen-2 imagery and edge-enhanced DeepLabV3+. The method includes: acquiring Gaofen-2 imagery and performing preprocessing such as radiometric calibration, atmospheric correction, geometric fine correction, and image fusion to construct a labeled training dataset; constructing an edge-enhanced DeepLabV3+ network, which adds an edge enhancement module to the standard DeepLabV3+ network to explicitly learn the edge features of winter wheat fields and inject edge information into the decoder; designing a joint loss function, including segmentation loss, edge loss, and boundary-aware loss, and training the network end-to-end; and using the trained network to perform sliding window prediction and post-processing on the image to be extracted to obtain the refined extraction results of winter wheat planting areas. This invention significantly improves the segmentation accuracy of winter wheat field boundaries through the edge enhancement module and boundary-aware loss.

This invention provides a method, system, and non-transitory computer-readable recording medium for projection image fusion correction. The method includes: after an electronic device is communicatively connected to multiple projectors, controlling each projector to project multiple corrected images onto a projection surface; generating captured images including imaging areas corresponding to the multiple corrected images using an image capture device; transmitting fusion information including the captured images to an artificial intelligence computing module for recognition, analysis, and processing to generate adjustment information; generating adjustment instructions based on the adjustment information and projector information; and transmitting the adjustment instructions to the corresponding projectors to adjust projection parameters. The projection image fusion correction method, system, and non-transitory computer-readable recording medium of this invention enable seamless fusion projection from multiple projectors.

The application discloses a ground glass nodule early lung adenocarcinoma invasiveness prediction method and system, comprising the following steps: first, image segmentation is performed on the preprocessed lung CT image through a three-dimensional image segmentation model to determine a lung lesion image of a lesion region in the lung CT image. Then, general features in the lung lesion image are captured through a three-dimensional image feature extraction model, and the general features are fused with the three-dimensional lung lesion image to obtain multi-modal image fusion features. Next, relevant text features are extracted from associated clinical text information through a text feature extraction model, and the image fusion features are fused with the text features. Finally, the fusion features are input into a pre-trained classifier to perform lung adenocarcinoma invasiveness prediction classification, so that the limitation of a single information source can be avoided, and the accuracy and reliability of ground glass nodule early lung adenocarcinoma invasiveness prediction can be significantly improved.

This invention discloses a CBCT multi-scan image fusion method and system. The method includes: controlling the scanning object to move along a preset scanning path and stopping at a reference scan position and multiple target scan positions respectively, synchronously acquiring the original CBCT image information and corresponding displacement information of the reference scan position and each of the target scan positions; reconstructing the original CBCT image information of the reference scan position and each of the target scan positions respectively to obtain initial three-dimensional CBCT images of the reference scan position and each of the target scan positions; aligning the initial three-dimensional CBCT images of each target scan position with the initial three-dimensional CBCT image of the reference scan position respectively to obtain aligned images; and using a multi-scale fusion algorithm to fuse the aligned images to obtain a three-dimensional CBCT stitched image. This invention solves the technical problems of low automation, low accuracy, and low efficiency in existing CBCT multi-scan image stitching.

The application discloses a molecular marker and MRI image fusion tumor early detection method and system, first collects serum molecular marker detection data and multi-sequence MRI image data of a digestive tract tumor screening object, constructs a molecular feature vector and extracts high-dimensional MRI image features respectively; a tumor prior probability is calculated through a molecular pre-screening deep neural network, and an adaptive spatial attention guide weight is generated; the weight is weighted and fused with the MRI image features to realize fine feature enhancement of a high-risk area and accurate lesion detection; based on a Bayesian evidence fusion reasoning framework, molecular detection evidence and image detection evidence are combined to update a posterior probability, and finally, a lesion positioning, detection confidence and visual detection report are output. The application realizes efficient complementation of molecular evidence and image evidence, greatly improves the accuracy, sensitivity and specificity of early detection of digestive tract tumors, and is suitable for non-invasive accurate screening of high-risk groups in clinical practice.

This invention provides a method and system for shallow water depth inversion based on multi-temporal satellite imagery guided by laser bathymetry. The method includes: extracting water body pixels from Sentinel-2 imagery; extracting water depth photons from satellite-borne laser bathymetry data and performing refraction correction to obtain water depth values ​​along the orbital direction; utilizing the strong correlation between the water depth values ​​along the orbital direction and the reflectivity of the blue and green bands of Sentinel-2 imagery to calculate multi-temporal image fusion weights and generate a weighted fused image; based on water column layer-by-layer inversion and / or seabed classification inversion strategies, dividing the weighted fused image into multiple homogeneous regions according to the water penetration capabilities of blue and green light; using LLM and / or LRM to invert water depth layer by layer in the multiple homogeneous regions; and fusing the results of different regions through a buffer zone to obtain the water depth inversion result for the entire shallow sea area. This invention can achieve high-precision and high-stability shallow water depth inversion.

This invention discloses an intelligent extraction and assessment method for ecological factors of musk deer habitat based on multi-source image fusion, belonging to the field of ecological remote sensing monitoring and intelligent assessment technology. It addresses the problems of inconsistent multi-source extraction of ecological factors and insufficient reliability of suitability assessment results for musk deer habitat. The method acquires optical remote sensing images, radar remote sensing images, UAV images, topographic data, meteorological data, vector geographic data, and ground survey data within the assessment area. After unified preprocessing, ecological factors such as vegetation, topography, water resources, climate, land cover, and human disturbance are extracted. Then, through multi-source fusion, sample matching, suitability probability calculation, key factor extraction, and result verification and correction, the suitability level of musk deer habitat and the contribution of ecological factors are obtained.

The application discloses a kind of based on multi-modal brain image fusion mood disorder identification method, belong to medical image processing and artificial intelligence auxiliary diagnosis technical field;Including the following steps: obtaining multi-modal brain image, registration to spherical space and alignment;After extracting and fusing features, spherical space-time self-attention coding and gate modulation are carried out;Modulated features are recursively down-sampled from initial level to penultimate level, and the aforementioned operation is repeated at each layer;In this level, features are processed in parallel, fused and compressed to the lowest level, while extracting clustering embedding features;Global features are aggregated at the lowest level, and after fusing with clustering embedding features, the recognition result is output by the clustering perception classifier;The application realizes multi-modal fusion and dynamic feature extraction under the maintenance of cerebral cortex topological structure, improves recognition accuracy and generalization ability.

This invention relates to the field of medical image 3D reconstruction and fusion technology, specifically a method and system for multimodal medical image fusion and 3D reconstruction, comprising: receiving and standardizing raw image data streams from different devices; extracting intensity and texture features of each modality of image, constructing a feature space mapping relationship, and generating a fused volume data field based on this using a feature-level fusion strategy; after filtering and enhancing the volume data, extracting isosurface geometric meshes and simplifying and smoothing them to obtain a preliminary 3D model surface; mapping the preliminary surface back to the volume data space, and performing surface detail optimization and topology correction; and finally outputting a 3D reconstructed model integrating complementary information from multiple modalities. This invention improves the information integration and geometric accuracy of the reconstructed model by performing deep fusion in the feature space and employing a closed-loop optimization mechanism of surface-volume data back-mapping.

This invention provides a topographic mapping method, medium, and system based on the fusion of laser point clouds and optical images, belonging to the field of surveying and mapping technology. This invention performs joint radiometric calibration of the reflection intensity of optical images and point clouds, outputting a radiometrically aligned image and a radiometrically calibrated point cloud. The above two types of data are input into a geometry-guided cross-modal dynamic graph attention fusion artificial intelligence model. Utilizing a differentiable K-nearest neighbor dynamic graph reconstruction module and an asymmetric bidirectional cross-attention mechanism, the geometric topological relationship of the point cloud actively guides image feature extraction, outputting dense point cloud semantic labels, surface normal vector estimation, occlusion confidence map, and texture quality assessment map. Based on the occlusion confidence map and texture quality assessment map, a dynamic task queue and work-stealing scheduling algorithm are used to generate topographic mapping results in parallel. This solves the problem of cross-modal feature alignment failure caused by the failure of geometric topological relationship to actively guide image feature extraction during cross-modal fusion of laser point clouds and optical images.

This invention discloses a method and system for identifying storm surge inundation ranges based on multi-source remote sensing imagery, particularly relating to the field of image analysis. The method includes: acquiring pre- and post-disaster images and digital elevation model (DEM) data within a target area; performing feature extraction and multi-source image fusion to generate a broad-spectrum suspected inundation area image; identifying complex underlying surface regions from the broad-spectrum suspected inundation area image based on pre-disaster optical images and obtaining corresponding binary mask images; performing texture feature analysis and pixel-level classification on each of the complex underlying surface regions to identify the actual water body regions; generating a preliminary inundation range map based on the actual water body regions of each complex underlying surface region; and performing terrain connectivity analysis on the preliminary inundation range map to generate a final storm surge inundation range map. This application can selectively filter out texture interference from different complex underlying surfaces in images based on pre- and post-disaster storm surge images, thereby improving the accuracy and reliability of storm surge range identification.

This invention discloses a registration and fusion system for heterogeneous multimodal medical images, belonging to the field of optoelectronic imaging and medical image processing technology. It includes: a structured light acquisition module for acquiring surface point cloud data of a target object; a hardware acceleration processing module comprising an FPGA device with a parallel pipelined registration logic embedded in the FPGA device, used to receive surface point cloud data and calculate its spatial transformation relationship with a preset reference point cloud; and an image fusion module for spatially aligning heterogeneous medical images with real-time images according to the spatial transformation relationship and generating a fused image. This application optimizes the data flow and computational pipeline through hardware and software co-design, reducing system latency, improving overall performance and stability, and facilitating clinical integration and application.

The application discloses a semantic segmentation method and device based on SAR image fusion edge perception function, and relates to the technical field of remote sensing image processing and computer vision. The method comprises the following steps: constructing a semantic segmentation network based on SAR images; extracting features of SAR images based on a CNN feature extraction encoder combined with a hollow convolution to obtain SAR image features; extracting features of edge contours between different subjects in the SAR images through an edge perception module to obtain features of edge contours between different subjects in the SAR images; calculating a semantic segmentation loss function of the network through a true value label and a probability distribution obtained by decoding SAR image features; calculating an edge segmentation loss function of the network through the true value label and a probability distribution obtained by decoding edge contour features; constructing a total loss function of the network according to the two loss functions; and training the total loss function to obtain a trained semantic segmentation network. The application can enhance the semantic segmentation interpretation effect.

This invention relates to the field of medical image detection technology, specifically to an intelligent lesion detection method based on multimodal medical image fusion. The method includes: acquiring multimodal medical image data of a target anatomical region; generating standardized image data blocks through cross-modal spatial registration and grayscale normalization preprocessing; performing multi-channel feature fusion enhancement on the image data, integrating texture, morphological, and functional metabolic features; performing three-dimensional semantic segmentation using a deep learning detection network to extract suspected lesion regions and obtain initial detection results; calculating multi-dimensional quantitative descriptors for each suspected lesion in the original image; and using a lesion classification and discrimination system to determine benignity / malignancy and pathological type. This method fully leverages the effective information from multimodal images, enriches the criteria for lesion discrimination, improves the precision of lesion identification, and accurately outputs results related to lesion localization, malignancy probability scoring, and pathological type.

The application discloses a thyroid cancer lateral neck lymph node metastasis endoscope-assisted modified Miccoli precise treatment system, and the system is operated through the following method, the method comprises the following steps: constructing a patient individualized three-dimensional anatomical model through preoperative multi-modal image fusion, and extracting the spatial topological relationship of the cervical sheath, accessory nerve, cervical plexus branch and lymphatic drainage path; planning a minimally invasive access path based on the three-dimensional anatomical model, and setting a double-channel puncture point at the posterior edge of the sternocleidomastoid muscle and 2 cm above the clavicle, which is respectively used for inserting a 30° oblique vision endoscope and a variable-angle ultrasonic aspirator; collecting tissue displacement data in real time through an electromagnetic navigation probe during the operation, and dynamically registering the preoperative model to generate a corrected navigation guide map; and the application aims to solve the problems of insufficient cleaning range, unclear anatomical level, limited operation space and disconnection between intraoperative real-time navigation and postoperative verification of the traditional Miccoli operation in the treatment of lateral neck (II-IV area) lymph node metastasis.

The application belongs to the technical field of image information, geographic information and AI model, and discloses a change monitoring method based on image fusion. The application collects image data by low-altitude collection, comprehensively uses technical means such as image data collection and management, image data fusion and reconstruction, plane projection model construction, intelligent data block division and coding, data block comparison and analysis, AI model detection and difference data geographic information retrieval, effectively solves the problems of existing technologies in aspects such as difficult data coupling, long data processing period, large result deviation and low data reusability, significantly improves data reusability, processing efficiency and standardization level, and forms efficient, flexible and intelligent image fusion change monitoring technology and application.