13 results about "Level fusion" patented technology

The invention discloses a multi-modal heterogeneous medical equipment data fusion and decision support method and device, and aims to solve the problems that the fusion precision is low due to space-time semantic difference of medical equipment multi-modal heterogeneous data (equipment operation parameters, clinical records, fault signals and the like), equipment management decisions depend on experience, and standards are not uniform. According to the method, breakthrough is achieved through three-level data alignment of'time-space-semantics', hierarchical fusion of'data level-feature level-decision level ', three-level decision driven by a knowledge graph and dynamic feedback optimization: time alignment uses a dynamic time warping algorithm, space alignment depends on a unified data dictionary, and semantic alignment introduces an attention mechanism; the feature level fusion quantifies the feature support degree based on the D-S evidence theory; the decision-making layer constructs a'rule-case-prediction 'three-level system, and combines cosine similarity retrieval and information entropy quantification uncertainty. The method and device can support medical equipment maintenance, clinical diagnosis and treatment and other scenes, and the medical service standardization level and the equipment management efficiency are improved.

The invention provides an identity authentication method which can be applied to the technical field of biological recognition. The identity authentication method comprises the following steps: after agreement or authorization of a user is obtained, collecting biological characteristic data of multiple modes of the user; preprocessing the collected biological characteristic data of various modes to generate corresponding biological characteristic vectors; performing quality evaluation on each biological feature vector to generate a corresponding quality score; dynamically calculating a weight coefficient of each biological feature vector in a feature fusion process by using a nonlinear weighting function based on the quality score; based on the weight coefficient, performing feature level fusion on each biological feature vector to generate a primary fusion feature; and performing cross-modal correlation analysis on the primary fusion features by using a multi-branch convolutional neural network based on an attention mechanism, and outputting an identity authentication result. The invention also provides an identity authentication device, equipment, a medium and a program product.

The invention discloses an intelligent artificial limb movement intention recognition method based on multi-modal information and a D-S evidence theory, and belongs to the technical field of artificial intelligence, mode recognition and intelligent artificial limb control. According to the method, firstly, electromyographic signals and kinematics signals of user limbs and three-dimensional point cloud information of the terrain where the user limbs are located are synchronously collected; filtering, segmenting and time domain / frequency domain feature extraction are carried out on the physiological signals, rotation correction and dimension reduction are carried out on the environment point cloud, and a binary image is generated; using a deep learning model to identify a preliminary motion mode from the physiological features and identify a terrain category from the environment image; and finally, inputting the two recognition results as evidence bodies into an improved D-S evidence theory model for decision-level fusion, and outputting a final comprehensive motion intention recognition result. According to the method, the human physiological signal and the external environment information are fused, so that the ambiguity and uncertainty of single information source identification are effectively solved, and the intention identification accuracy and adaptability of the intelligent artificial limb in different complex scenes are remarkably improved.

The invention discloses an intelligent tumor diagnosis large model construction method based on multi-modal medical data fusion. The method comprises the steps of 1, data collection processing and feature extraction, 2, establishment of a dynamic fusion model, 3, arrangement of a training set, 4, model training optimization, 5, establishment of an auxiliary diagnosis model and 6, model deployment. According to the method, a three-level fusion mechanism of a self-adaptive fusion framework is established, medical image data and clinical data of various modes are effectively fused, the model is trained by using a double-domain transfer learning strategy, the small sample learning efficiency and precision are improved, the characteristics of tumors can be reflected more comprehensively and accurately, and the method is suitable for large-scale popularization and application. A powerful auxiliary reference is provided for improving the accuracy and efficiency of tumor diagnosis, and the model can be expanded and deployed in other disease auxiliary diagnosis reference systems and has a wide application prospect.

The invention provides a mechanical response field reconstruction method based on multi-source data multilayer fusion, and relates to the technical field of modern industry and engineering, and the method comprises the steps: S1, obtaining multi-source data of a mechanical response field; wherein the multi-source data comprises local high-precision strain field data, global finite element field data and high-precision strain gauge measurement data; performing coordinate transformation on the multi-source data to obtain a unified data coordinate; based on the unified data coordinates, performing first-level data fusion on the local high-precision strain field data and the global finite element field data to obtain a first-level fusion field; and performing secondary data fusion on the first-level fusion field and high-precision strain gauge measurement data, performing analysis by using a position prediction model to obtain a second-level fusion field, and completing reconstruction of the mechanical response field. The problems of how to accurately fuse multi-type data and how to improve the accuracy and reliability of large-scale structure test monitoring are solved.

The application provides a lumen surgery instrument action classification method based on kinematic appearance fusion and related devices. The method comprises the following steps: S1. acquiring a lumen surgery video frame sequence and an instrument tip motion trajectory; S2. performing appearance feature extraction on the lumen surgery video frame sequence to obtain an appearance feature vector; S3. performing kinematic feature extraction on the instrument tip motion trajectory to obtain a kinematic feature vector; S4. performing feature-level fusion on the appearance feature vector and the kinematic feature vector to obtain a fusion feature vector; S5. performing classification prediction based on the fusion feature vector to obtain an action category probability distribution; and S6. performing time sequence smoothing on the action category probability distribution based on a hidden Markov model to output a final instrument action category and a confidence degree. The application also provides related devices corresponding to the method, and the related devices comprise a device, an electronic device, a computer readable storage medium and a computer program product.

The application discloses a human posture estimation and behavior analysis method, comprising the following steps: S1, collecting visible light images and thermal imaging images, performing feature extraction and fusion, and generating fused semantic features; S2, obtaining a refined key point heat map and a partial affinity field according to the fused semantic features; S3, generating a key point set according to the refined key point heat map and the partial affinity field; and S4, determining an abnormal behavior according to the key point set. The application effectively overcomes the shortcomings of large pixel-level fusion calculation overhead and difficult alignment of shallow feature fusion modalities, and significantly improves the robustness and precision of human posture estimation under complex lighting conditions such as day-night alternation and backlight.

This invention discloses a deep learning-based InSAR geological hazard deformation identification method, belonging to the field of geological hazard monitoring and prevention technology. The method includes the following steps: S1, multi-source data acquisition; S2, establishing a multi-source data spatiotemporal registration and pixel-level fusion model; S3, constructing a joint extraction network for multi-scale three-dimensional deformation features; S4, introducing elastic mechanical constraints to optimize the three-dimensional deformation field inversion; S5, fusing the physical model and deep learning output uncertainty quantification results, finally outputting the three-dimensional deformation of each pixel and its corresponding confidence assessment, forming a quantitative deformation result map for geological hazard risk assessment. This deep learning-based InSAR geological hazard deformation identification method employs surface deformation monitoring via synthetic aperture radar interferometry, multi-source remote sensing data fusion, and intelligent deformation inversion and risk assessment combining deep learning and physical constraints. It can be applied to the identification, monitoring, and risk assessment of geological hazards.

The application discloses a two-dimensional human body posture estimation method based on a multi-level fusion and discrimination network, which firstly extracts texture information of an image through a high-resolution CNN network, then improves the feature expression capability of the network by fusing multi-level features, and then positions a rough human body contour through a discrimination network, and then is sent into an efficient Transform module for processing; in different stages of the Transform, confidence scores of each Token are scored, and the score size determines the importance of the Token relative to a skeleton key point, and Tokens with low importance are fused into a new Token to reduce information redundancy and improve the calculation efficiency. Finally, the application is tested on mainstream COCO and MPII data sets, and is superior to mainstream most advanced models in terms of calculation complexity and network accuracy.

The invention relates to the technical field of medical artificial intelligence and high-performance computing, in particular to a mercuric chloride platform-based lightweight multi-modal medical imaging intelligent diagnosis method, which comprises the following steps of S1, receiving and standardizing multi-modal medical data through a preprocessing module; s2, the standardized multi-modal data is mapped to a unified semantic space through a cross-modal semantic fusion module, feature level fusion is carried out, and deep semantic representation is generated; s3, through a cloud-edge collaborative reasoning module, according to the complexity of a diagnosis task, calling a diagnosis model at a cloud end or an edge end to perform reasoning on the deep semantic representation; and S4, performing multi-scale enhancement on the features in the reasoning process through a multi-scale focus recognition module. According to the diagnosis comprehensiveness breakthrough, deep collaborative reasoning of multi-source heterogeneous data is achieved through the cross-modal unified semantic mapping technology, and the comprehensive diagnosis index MMMU-Med reaches 82.3 scores and is far better than that of a single-modal model.

The invention relates to the technical field of image analysis, and discloses a multi-modal image fusion cylindrical skeleton repair detection method and system, and the method comprises the steps: collecting the multi-modal image data of a cylindrical skeleton, recognizing the bone trabecula structure feature points of each modal in the multi-modal image data, carrying out the modal space registration of the multi-modal image data, and carrying out the recognition of the feature points of the bone trabecula structure. A spatial registration image is obtained; using the spatial registration image to identify a multi-dimensional cylindrical skeleton feature of the cylindrical skeleton; performing feature level fusion on the multi-dimensional cylindrical skeleton features to obtain a fusion feature map of the cylindrical skeleton, and constructing a skeleton three-dimensional model of the cylindrical skeleton by using the fusion feature map; and utilizing the three-dimensional skeleton model to identify the callus volume and the bone cortex continuous state of a repair area corresponding to the cylindrical skeleton, analyzing the skeleton integrity and the adjacent bone fusion degree of the repair area, and constructing a repair detection report of the cylindrical skeleton. According to the invention, the precision of cylindrical skeleton repair detection can be improved.

The application discloses a carotid plaque tracking and echo classification framework and method of an ultrasonic video, a multiscale hollow encoder, an internal and external feature decoupler and a tracking network are combined to form a carotid plaque tracking framework, and a carotid plaque echo classification framework is established based on feature recombination and a double-channel 3D-Attention framework; an ultrasonic video image is input into the tracking framework to position a plaque position, feature recombination is used to select a plaque contour as local feature input of the echo classification framework, and an original ultrasonic video image is used as global feature input of the echo classification framework. The tracking and classification framework combines deep learning technologies such as feature hierarchical level fusion, multiscale context temporal feature extraction and a three-dimensional attention mechanism, not only solves the optimization problem caused by the change of the state between ultrasonic sections in the carotid ultrasonic video, and fully combines the context features and the vascular environment features of the plaque, so that the features corresponding to each type of state reflecting the instability of the plaque are fully reflected.

The invention belongs to the technical field of behavior recognition, and provides a skiing level evaluation method and system based on multi-source information fusion, and the method comprises the steps: synchronously collecting multi-source time sequence data and biological information when a skier completes a preset key action, and carrying out the standardization processing; constructing a data set, wherein the data set comprises a time sequence data set composed of multi-source time sequence data and a biological data set composed of biological information; constructing a multi-task neural network model, and synchronously completing a behavior recognition task and an identity recognition task; training the multi-task neural network model by using a training set containing excellent athlete sample information; utilizing the trained multi-task neural network model to extract high-level fusion features of a to-be-evaluated skier; and calculating the similarity between the high-level fusion features of the to-be-evaluated skier and a preset excellent skier, and evaluating the skiing level according to the similarity. According to the method, the problem that an evaluation system in the prior art cannot fully consider individual differences among skiers is solved.