80 results about "Connectome" patented technology

The invention relates to a multistable mechanical metamaterial structure with a tension characteristic, which belongs to the field of mechanical metamaterials and mainly comprises a multistable mechanical metamaterial cell structure, and the multistable mechanical metamaterial cell structure comprises a tension integral structure component, a degree-of-freedom constraint component and a connecting component. Under the action of a tensioning integral structure, when the structure is not subjected to external load, the structure is subjected to interaction of internal stress, and the balance is achieved in a self-stabilization state; according to the multistable mechanical metamaterial cell element structure with the tension characteristic, the structure of the position where each side of the square outline is located has two steady-state configurations, and the cell element structures are connected through a rectangular array to form the multistable mechanical metamaterial structure with the tension characteristic. The tensegrity structure and the mechanical metamaterial structure are combined, the types of existing mechanical metamaterial cell element structures are enriched, and a brand new direction is opened up for design, analysis and application of the tensegrity mechanical metamaterial structure.

The invention discloses a high-precision multi-physics field reconstruction method based on a nested neural network architecture. The method comprises the following steps: S1, acquiring open-source multi-physics field data; s2, a multi-scale Fourier operator-driven U-Net physical field reconstruction model is constructed, the physical field reconstruction model comprises an embedding module, a coding module and a decoding module, the coding module comprises at least one coding layer, each coding layer is formed by sequentially connecting a multi-scale Fourier neural operator layer and a down-sampling layer, and the multi-scale Fourier neural operator layer and the down-sampling layer are embedded into the embedding module; the decoding module comprises at least one decoding layer, and each decoding layer is formed by sequentially connecting a multi-scale Fourier neural operator layer and an up-sampling layer; s3, training the physical field reconstruction model based on a training set; and S4, performing quantitative analysis and qualitative analysis on the reconstructed physical field by using the test set.

Example implementations relating to mounting assemblies are disclosed herein, including in one example a skull mount having a substantially longitudinal bracing member and an engaging member comprising a curved portion configured to engage a portion of an animal skull such as a base of the brain cavity. One or more adjusting members adjustably position and secure the bracing member and the engaging member to securely engage and retain a skull of an animal therebetween by engaging the curved portion of the engaging member with at least a portion of the skull such as the brain cavity. A base assembly attached to a mounting surface is coupled to a connecting assembly removably coupling the bracing member and engaging member assembly to the base assembly, the connecting assembly comprising a pivot rod and a sleeve configured to receive and removably retain the pivot rod.

The invention relates to the technical field of brain science image processing, and discloses a brain tissue neuronal morphology three-dimensional reconstruction and connection group analysis method and system, and the method comprises the steps: carrying out the feature point extraction and matching and optical flow registration of continuous ultrathin section electron microscope images, and obtaining three-dimensional image volume data; carrying out semantic segmentation on cell bodies, dendrites, axons and dendritic spines through a three-dimensional deep convolutional neural network, and driving secondary registration through confidence feedback; detecting synapses and classifying the synapses as excitatory or inhibitory types; executing topology perception skeletonized reconstruction and performing segmentation correction through topology anomaly feedback constraint; and constructing a synaptic connection matrix and executing graph theory analysis.

The invention provides an intelligent process arrangement method, system and device based on vue-flow and a medium, and belongs to the technical field of front-end process arrangement. A node data structure and an interaction rule are unified by formulating a node standardization interface and an attribute panel interface; developing a public connection component and providing a style self-definition capability; packaging special nodes such as a large model based on a standardized interface; constructing a node registration class to manage node uniqueness and connection logic, and providing a component registration interface and a development auxiliary method; the registration node is adapted to a vue-flow framework, and normal rendering is ensured; and finally realizing the rendering logic and state style unification of the node and the attribute panel. The standardization degree and the development efficiency of a process arrangement system are improved, the node compatibility is guaranteed through a standardized interface, and the framework adaptation difficulty is reduced. Connection logic management and state styles uniformly optimize user experience; according to the overall scheme, the stability and maintainability of the system are enhanced, and a solid technical support is provided for intelligent process arrangement.

The application discloses a physiological signal monitoring method and a wearable device, relates to the technical field of detection, and mainly aims to solve the problem of poor accuracy of existing electrocardiogram for heart state judgment. The method comprises the following steps: acquiring user basic information and a physiological signal collected in real time, wherein the physiological signal comprises a heart electrocardiogram signal, a muscle electrocardiogram signal and an acceleration signal, the physiological signal is collected based on a multi-modal sensor on a wearable device, and the wearable device is fixed on the user body through a belt type connection assembly; performing multi-modal fusion on the physiological signal based on a multi-modal fusion algorithm to obtain physiological parameter data, and determining a monitoring task, wherein the multi-modal fusion algorithm comprises one of parallel fusion, serial fusion and attention fusion; calling a monitoring prediction model matched with the monitoring task and having completed model training to perform monitoring processing on the physiological parameter and the user basic information, obtaining a monitoring result, and performing feature comparison based on the monitoring result.

The invention discloses a multi-modal medical image segmentation system based on modal complementation and state space modeling, and the system is constructed based on a 3D U-Net architecture, and comprises a first residual module, a second residual module, a first connection module and a segmentation head which are connected in sequence, N encoders which are formed by connecting convolution modules and down-sampling modules in sequence, each decoder is formed by sequentially connecting a residual error module, a connection module and a transposition module. One path of 3D input data is input to the depth channel convolution module, and the other path of 3D input data is input to the first residual module. And a transpose module output signal of the Nth decoder and a second residual module output signal are input into the first connection module together, are connected by the first connection module and then are output into the segmentation head, and the segmentation head outputs a segmentation signal of the multi-modal data. According to the method, the complementary characteristic between different modes in the multi-mode medical image is fully utilized, and the global context modeling and local structure sensing capabilities are considered at the same time.

The application discloses a robot positioning method and system based on deep learning, relates to the technical field of artificial intelligence, and sends a global environment map and a local environment map under the global environment map into a trained deep neural network model to obtain an accurate candidate pose of a robot; the candidate pose is finely corrected through robot movement, and finally the best robot pose is obtained; the deep neural network model comprises a convolution layer conv0, a maximum value pooling layer maxpool, a convolution layer conv, a cross attention layer cross_attn and a candidate pose acquisition layer Maxop; the maximum value pooling layer maxpool and the convolution layer conv are sequentially connected to form a layer group, a plurality of layer groups are sequentially connected, the layer group located at the uppermost layer is connected to the output of the convolution layer conv0 as the shallowest scale, and the output of the convolution layer conv0 and all layer groups are sequentially connected to the cross attention layer cross_attn and the candidate pose acquisition layer Maxop to output the best pose of the robot; the robot positioning method and system improve the accuracy and robustness of robot positioning.

The application discloses a magnetocardiogram gradient instrument and an electrocardio collection circuit packaging structure and a signal processing system, relates to the technical field of medical equipment, and realizes high-density integration of hardware and reliable protection through a rigid protection frame formed by a circuit chip, a layer plate structure and a connecting assembly, shortens a signal transmission path, reduces the influence of environmental vibration, dust, moisture and electromagnetic disturbance on a synchronous sampling link, and provides a stable data basis for multi-modal fusion imaging. The signal processing system comprises multi-modal synchronous collection, noise suppression, data fusion imaging and a diagnosis auxiliary module, generates a high-precision space-time dynamic imaging atlas by constructing a cardiac electromagnetic field joint forward model and applying time and space double constraints, and outputs a cardiovascular risk assessment, a cardiac arrhythmia positioning and a myocardial ischemia quantification result in combination with machine learning. The application solves the problems of low integration, poor anti-interference and insufficient single modal detection precision of existing equipment.

The embodiment of the invention discloses a testing device. The testing device comprises a cable, a connecting assembly and a stimulating assembly which are connected in sequence. Wherein the connecting assembly is provided with a limiting groove for limiting and fixing the implantable electrode, a plurality of conductive elastic sheets in conductive connection with the cable are arranged in the limiting groove, and the conductive elastic sheets are used for being detachably and electrically connected with a wire of the implantable electrode, so that the implanted implantable electrode is tested. The stimulation assembly is provided with a conductive probe electrically connected with the cable, and can apply a test electric signal to the nervous tissue to carry out a stimulation reaction test, thereby assisting in accurately confirming the position of the target spot. According to the testing device, the electrode testing function and the nerve stimulation testing function are integrated into a single device, the structure is compact, operation is visual, and the accuracy, smoothness and convenience of an operation are remarkably improved.

The utility model discloses an acquisition / stimulation device and a dual-mode electromyography, and relates to the electromyography technical field, the acquisition / stimulation device comprises a housing, a PCB, a stimulation assembly, an acquisition assembly, a control assembly and a connection assembly, the housing is provided with an accommodation cavity; the PCB is arranged in the accommodating cavity; the stimulation assembly is arranged on the PCB; the acquisition assembly is arranged on the PCB; the control assembly is arranged on the PCB and electrically connected with the stimulation assembly and the acquisition assembly. The connecting assembly is arranged on the PCB, one side of the connecting assembly is electrically connected with the control assembly, and the other side of the connecting assembly is partially exposed out of the shell and is used for being electrically connected with the electrode plate; the control assembly controls the connection assembly to be electrically connected with the stimulation assembly or the acquisition assembly, so that the acquisition / stimulation device has an acquisition state and a stimulation state. The acquisition / stimulation device provided by the utility model aims to realize two functions of acquiring electromyographic signals and improving or recovering stimulated muscles or muscle groups through one set of equipment.

This application provides a disease intelligent diagnostic device based on brain structural connectivity identifiers and its application. The device includes (1) a data acquisition module; (2) a data processing module: reconstructing the spin distribution function of the data and distributing the spin distribution function in a standard space; (3) a data projection module: projecting the spin distribution functions of patients and healthy individuals into the standard space to obtain the Z-value of the maximum direction within the voxel; and (4) a Connectome Identifier extraction module: reducing the dimensionality of the Z-value to form a 1-dimensional feature value. The device of this application can be used for the diagnosis of various brain diseases and the determination of lesion location and severity.

The invention discloses a mental disease typing method and system integrating DNA methylation and neuroimaging, and the method is a neuroimaging biological annotation method integrating DNA methylation and brain connection group data, and comprises the steps: recognizing brain network features related to a specific molecular phenotype in a whole brain range through a machine learning model; and mechanism-sensitive layering of the mental disorder heterogeneity group is realized. The method does not need to depend on a prior classification or hypothesis mechanism, can be suitable for different types of mental disorder people, provides technical support for exploring potential biological mechanisms and identifying targeted therapy groups, and has high generalizability and clinical application prospects.

This invention provides a highly supportive mesh structure, comprising beam-column units and single expansion units. The beam-column units at both ends and the single expansion unit in the middle are connected to form a tubular mesh structure. Each single expansion unit is composed of single expansion grids axially connected by reinforcing bars, with the crests of radially adjacent single expansion grids arranged in an alternating pattern. This mesh structure exhibits strong radial support, low axial contraction, no sharp angles, and stable mechanical properties. It is adaptable to various cavities in the human body, enabling the establishment and maintenance of cavities blocked by various hard, sharp, and irregularly shaped foreign objects. This provides surgeons with a safer surgical procedure, increases their treatment options, and reduces patient pain and suffering during and after surgery.

The utility model relates to the technical field of hand multi-joint collaborative rehabilitation, in particular to a hand multi-joint collaborative rehabilitation trainer which comprises a base, a roller is arranged at the top of the base through a connecting assembly, a driving block is transversely arranged in the center of the interior of the roller, and the five faces of the roller are each connected with a set of training rods in a sliding mode. Each group comprises five training rods, five electromagnets are fixedly installed on the top of the driving block, the five electromagnets correspond to the five training rods of each group respectively, and the five electromagnets are powered by independent loops respectively. The device has the advantages that the magnetic force is adjusted by changing the current intensity of the electromagnet, then the force required by a patient to pull the training rod is changed, difficulty is gradually increased along with the recovery progress of the patient, personalized training is achieved, training singleness is avoided, the training rod is randomly attracted to descend, and the patient pulls the training rod back to the original position as soon as possible; the hand muscle strength of the patient is exercised, and the brain reaction speed and the hand-brain coordination ability of the patient are improved.

This disclosure generally relates to a computer-implemented method for isolating the physical locations of components within a physical structure. This is consistent with the present disclosure, which involves performing a series of evaluations at evaluation points depending on the type of physical structure and combining the results of the performed evaluations. The solution according to this disclosure can be used in many different technical fields, for example, for identifying erroneous locations within connection components, such as wire harnesses. This disclosure also relates to corresponding computer systems and computer program products.

The invention provides an animal sleep control device and a sleep control method.The animal sleep control device comprises a wearing part, a first collector and a second collector, the first collector and the second collector are arranged on the wearing part, the wearing part is of a fan-shaped annular structure when unfolded, connecting assemblies are arranged on the two sides of the wearing part in the circumferential direction respectively, and when the connecting assemblies are connected, the first collector and the second collector are connected. The wearing piece encloses and forms a wearing structure with an external expansion shape; the wearing part comprises a control assembly located on the enclosed inner surface, a main control assembly is arranged on the outer surface of the wearing part, the control assembly, the first collector and the second collector are all connected with the main control assembly, and the first collector and the second collector collect signals and transmit the signals to the main control assembly; the main control assembly analyzes the signal and drives the control assembly to work according to the signal; according to the application, through arrangement and cooperation of various control elements, gradual enhancement of animal sleep deprivation stimulation is realized to enhance the sleep deprivation effect, so that the accuracy and authenticity of experimental data acquired by the remote control terminal are improved.

The invention discloses an image encoding and decoding method based on residual space Mama. Nonlinear transformation modules are inserted between adjacent down-sampling operations of an encoder of an encoding and decoding system and between down-sampling operations of a decoder of the encoding and decoding system; the nonlinear transformation module is a residual space Mama module or a mixed attention module, the residual space Mama module is formed by sequentially connecting a first LayerNorm layer, a structure sensing state space module, a second LayerNorm layer and a feedforward neural network, and jump residual connection is carried out at output; the hybrid attention module introduces a block-based channel attention mechanism module to the output of a feedforward neural network of the residual space Mamba module; according to the method, local and global spatial context information of the image can be effectively captured, and the calculation complexity is reduced to linear growth on the basis of ensuring adaptive weight adjustment; and in combination with a block-level channel attention mechanism, fine-grained modeling is performed on redundancy characteristics in the channel while global spatial information is captured, so that the compression capacity of a potential variable space is improved.

The application is suitable for the technical field of surgical robots, and provides a surgical robot, a medical instrument withdrawal method and a readable storage medium. The surgical robot comprises a puncture device configured to be connected to a living body, a medical instrument comprising a connecting assembly, the connecting assembly comprising a first joint, the connecting assembly being inserted into the living body through the puncture device, and a controller coupled to the medical instrument and configured to: in response to a request for withdrawing the medical instrument, determine an instruction set, the instruction set comprising a first instruction for commanding the first joint to recover from a current state to an extended state and a second instruction for commanding the medical instrument to be axially retracted by a target distance along the puncture device; control the first joint to be extended one by one according to a configured order according to the first instruction, and control the medical instrument to be axially retracted by the target distance along the puncture device according to the second instruction, so that the connecting assembly is extended and retracted to a target position of the puncture device, and the medical instrument is flexibly and safely withdrawn.

The invention discloses a three-dimensional visual modeling method for a docking mechanism envelope space. The three-dimensional visual modeling method comprises the following steps: defining a characteristic coordinate system and a nominal position type of a docking mechanism; acquiring an envelope space boundary condition of the docking mechanism; simplifying geometric envelopes of the docking mechanism and a connecting assembly thereof, and establishing a virtual prototype standard envelope model of the docking mechanism; according to the virtual prototype standard envelope model of the docking mechanism and the boundary conditions of the envelope space, establishing a three-dimensional visual virtual prototype simulation model of the envelope space of the docking mechanism; and outputting a three-dimensional visual virtual prototype simulation model of the docking mechanism envelope space. According to the three-dimensional visual modeling method for the envelope space of the docking mechanism, the problems that in the prior art, a data description mode of the envelope space of the docking mechanism is not visual, and a visual modeling method for the envelope space of the docking mechanism is lacked are solved.

The invention discloses a head-mounted non-invasive intracranial pressure monitor adaptive to a skull flap-removed patient, and belongs to the technical field of neurosurgery monitoring equipment.The head-mounted non-invasive intracranial pressure monitor comprises a head-mounted structure, the head-mounted structure comprises a center plate and a plurality of clamping strips connected to the center plate, and each clamping strip is provided with a plurality of detection assemblies; the detection assembly is composed of a telescopic pipe and an electrode patch. A regulation and control box is arranged on the head-mounted structure, a connecting assembly is arranged between the regulation and control box and the head-mounted structure, an air control box is arranged in the regulation and control box, an air outlet is formed in one side of the air control box, and the air outlet is connected with the telescopic pipes through hoses. By means of structural innovation and intelligent control, noninvasive, accurate and safe monitoring of the intracranial pressure of the patient without the bone flap is achieved, and a brand new solution is provided for neurosurgery postoperative monitoring.

The application is a kind of bionic nerve circuit model based on artificial synapse device. The circuit model is composed of n multi-stage connections of synapse-LIF artificial neuron devices. The composition of the synapse-LIF artificial neuron device includes synapse-like device and cytoplasm-like device. The synapse-like device includes artificial synapse device, voltage dividing resistor and voltage control switch. The cytoplasm-like device includes one comparator. Through different combinations of the four devices, namely excitatory synapse-like device, inhibitory synapse-like device, excitatory cytoplasm-like device and inhibitory cytoplasm-like device, the devices of six circuits, including feedforward excitation circuit, feedforward inhibition circuit, feedback inhibition circuit, lateral inhibition circuit, disinhibition circuit and mutual inhibition circuit, are formed. The application realizes the functions of perception afferent, information integration and motor control in the motor nervous system through the circuit model group, which is helpful for the repair or reconstruction of damaged motor conduction circuit.

This invention discloses an abnormal heart rate diagnostic system based on a lightweight spiking neural network, comprising: a data denoising module for denoising the input electrocardiogram (ECG) signal; a data segmentation module for segmenting the denoised ECG signal; a network construction module for constructing a lightweight spiking neural network consisting of a feature extraction network and a classification network connected sequentially; wherein the feature extraction network includes multiple branches of feature extraction modules, each branch including several sequentially connected IFR modules; the classification network includes first and second layer SGRU modules connected sequentially; a network training module for training the constructed lightweight spiking neural network using the segmented ECG signal to obtain a trained lightweight spiking neural network; and a data diagnosis module for inputting the ECG signal to be diagnosed into the trained lightweight spiking neural network to obtain a diagnostic result. This invention improves the accuracy of ECG signal classification.

The application provides a vue-flow-based intelligent process arrangement method, system, device and medium, belongs to the technical field of front-end process arrangement, and through formulating a node standardized interface and an attribute panel interface, unifying node data structure and interaction rules, developing a public connection component and providing style customization capability, encapsulating special nodes such as large models based on the standardized interface, constructing node registration class management node uniqueness and connection logic, providing component registration interface and development auxiliary method, adapting the registered node to the vue-flow framework, ensuring normal rendering, and finally realizing the rendering logic and state style unification of the node and the attribute panel. The standardization degree and development efficiency of the process arrangement system are improved, the node compatibility is ensured through the standardized interface, and the framework adaptation difficulty is reduced. The connection logic management and state style unification optimize the user experience; the overall scheme enhances the system stability and maintainability, and provides solid technical support for intelligent process arrangement.

The invention discloses a human behavior recognition method and device based on multi-scale adaptive graph convolution, and the method comprises the steps: S1, obtaining joint data, skeleton data and motion data according to 3D human motion data; s2, respectively inputting the joint data, the skeleton data and the motion data into three corresponding information flow sub-networks; wherein each sub-network is respectively stacked with ten multi-scale adaptive large kernel graph convolution modules, and each multi-scale adaptive large kernel graph convolution module is formed by connecting a space module, a time module and a residual error; and S3, performing classification operation according to the feature data output by the three information flow sub-networks to realize human body behavior recognition. By adopting the technical scheme of the invention, global and local motion features are effectively aggregated, and the expression ability of the model is improved while the calculation efficiency of the model is kept.

The invention provides a cerebral apoplexy postoperative care risk early warning method and system fused with a brain image, and the method comprises the steps: improving the prediction precision through multi-modal data fusion, and integrating clinical physiological features, focus function influence points and brain connection group disconnection features to construct an initial data set; learning the importance of each feature by using the primary model, adjusting the weight of the brain function map, and optimizing the focus influence integral to enhance the representation ability; for a sample imbalance problem, generating a synthetic sample based on an importance coefficient according to a local density and a weighted distance, and realizing minority class expansion; and training a risk early warning model based on the balanced data set, inputting a new patient by adopting the same feature extraction process, outputting a personalized postoperative care risk level, and realizing an intelligent assessment closed loop from image anatomy to a functional network.

The invention relates to the technical field of intelligent recommendation, in particular to a method, device and equipment for recommending an anchor sequence for a user and a medium, the method comprises the steps that user information, anchor information and associated context information between the user and an anchor are collected, and the user information comprises sequence information that the user watches historical anchors and basic information of the user; preprocessing the user basic information, the anchor information and the context information to obtain grouping feature information; on the basis of the grouping feature information, the grouping feature information is mapped to Embedding vectors respectively; generating a user self-attention vector based on the sequence information that the user watches the historical anchor; based on the two, obtaining a mixed cross vector; based on a hybrid cross vector, an Embedding vector and a user self-attention vector, in a recommendation model, different deep neural network networking is constructed, a full-connection networking for different behavior data of the user is constructed in a hybrid manner, an anchor sorting result is recommended for the user, the model complexity is effectively reduced, and the estimation overhead is reduced.

The invention provides an emergency communication spectrum prediction method based on a gating adaptive time sequence diagram integrated network, which relates to the technical field of communication, and comprises the following steps: acquiring multi-band historical spectrum data, and dividing a training sample and a prediction label according to a time step; constructing a self-adaptive dynamic graph structure fusing priori knowledge and a learnable matrix, and combining gating time sequence convolution, graph convolution, channel alignment convolution and residual connection to form a gating time sequence graph network; carrying out iterative training by utilizing the training sample and the dynamic graph structure, and obtaining an integrated prediction model through period storage and average weight snapshot; and inputting a to-be-tested sample into the model and outputting a frequency spectrum prediction value. The method has the advantages that spatio-temporal feature extraction is fused, and prediction precision and scene adaptability are improved; the generalization and robustness of the model in the environment change are enhanced by the self-adaptive dynamic graph structure; the parallel gating time sequence convolution is adopted to reduce the calculation complexity and the reasoning time delay, and the millisecond-level response requirement of emergency communication is met.

The invention relates to an electroencephalogram traceability model training method and device, equipment and a medium. The method comprises the following steps: constructing a standardized graph structure data set containing multiple individual electroencephalogram signals, a structure connection group and a source activity true value, and training a graph neural network by adopting a meta-learning framework to extract a common rule of a cross-individual brain connection group and a traceability mapping relationship, so as to form a pre-training model with strong generalization ability; for a new individual, only key parameters associated with connection group features in the model are adjusted through a parameter efficient fine tuning technology, and rapid migration of pre-training meta-knowledge to individual specific connection is realized; finally, while individualized traceability precision is kept, computing resources and data volume required for model adaptation are greatly reduced, the problem of traceability deviation caused by individual brain structure difference in a traditional method is effectively solved, and feasibility and efficiency of an electroencephalogram traceability technology in clinical practice are remarkably improved.

The invention discloses a data processing method and device, equipment and a storage medium. Comprising the steps that a target relation graph is obtained, and the target relation graph is formed by connecting a plurality of relation nodes and edges between the relation nodes; searching an intra-graph neighbor node of each relation node in the target relation graph; based on each relation node and the intra-graph neighbor node corresponding to each relation node, determining the intra-graph neighbor relation of each relation node in the target relation graph; according to the intra-graph neighbor relation of each relation node, calculating the node relation importance degree of each relation node, and according to the shortest path of each relation node pair in the target relation graph, calculating the node position importance degree of each relation node; and based on the node relation importance and the node position importance corresponding to each relation node, determining a target key node in the target relation graph. According to the method, the recognition speed and expansibility of the key nodes in different application scenes are improved.