14 results about "Lateral inhibition" patented technology

The invention relates to the technical field of artificial intelligence, can be applied to business scenes such as agent autonomous decision making, financial science and technology and medical health, and discloses a task processing method and device based on visual attention enhancement, equipment and a medium. Visual hierarchical features are extracted, a double fovea attention module processes and fuses high-level visual features, a side suppression network obtains enhanced visual features, and a cross-modal fusion module generates fusion features by taking the enhanced visual features as query vectors and taking language components and action components as key and value vectors; and fusing the feature input decision network to generate target category and position information, generating feedback information based on actual label difference, and updating module parameters to complete a target task. According to the invention, through combination of a bionic vision mechanism and multi-modal attention fusion, the visual feature extraction and background suppression capability is improved, and the target capture efficiency and recognition precision in a complex scene can be improved.

The invention belongs to the technical field of biological information calculation, and discloses a pet feed additive formula calculation model optimization system, which comprises a data interface for acquiring raw material parameters and data containing environmental stress factors; the memory stores a heterogeneous neural network model, and comprises a first time sequence convolution subnet, a second cascade processing subnet and a global gating distribution unit which are arranged in parallel and are coupled through a lateral suppression channel; the processor operates the model, the global gating distribution unit is used for dynamically adjusting the computing power weight between the subnets according to the system energy state tensor, and back propagation iteration is carried out based on a nonlinear constraint penalty term. The problems of convergence difficulty and negative migration when a general model processes high-dimensional strong-coupling biological data are solved, and the prediction precision and safety of formula optimization are improved.

The invention discloses an MI-EEG classification and recognition method based on a pulse neural network. The method comprises the steps that electroencephalogram signals are collected and preprocessed; converting the two-dimensional time-frequency image into a pulse sequence in a Poisson coding mode; generating a frequency histogram of excitation neuron pulse distribution by inputting the pulse sequence into a pulse neural network; and classifying the output pulses by adopting a voting method. According to the invention, the pulse neural network is used to identify and classify the image, so that the precision meets the requirement, and the expenditure of power consumption and the like is reduced; meanwhile, an STDP learning mechanism and a lateral inhibition mechanism are introduced, the connection strength is adjusted through the STDP learning mechanism, and excitation neurons of unissued pulses are inhibited through the lateral inhibition mechanism; the STDP learning mechanism and the side suppression mechanism jointly influence the neuron group, so that the neurons of the corresponding instructions can emit pulses more easily, the membrane potential of the inactive neurons is reduced, the cost of emitting the pulses is increased, and the pulses are less likely to be emitted.

The application discloses a molecular property prediction method based on spectral position coding and bionic side inhibition gating, and relates to the field of olfactory perception. The method comprises the following steps: constructing a molecular structured input comprising atomic types, a Coulomb matrix and graph Laplacian eigenvector; fusing atomic chemical attributes and spectral domain topological positions through a feature extraction module to generate initial atomic features; using a multi-scale aggregation module, hierarchical neighborhood aggregation is carried out based on preset scale constraints to extract multi-level structure features covering chemical bonds, functional groups and molecular skeletons; with the help of a global interaction module of bionic side inhibition gating, local features are dynamically modulated by a global query signal to realize adaptive denoising and semantic sharpening; finally, through mask pooling and a decoupled multi-label prediction head, the scores of each odor attribute of the molecule are output. The application significantly improves the accuracy, structure perception ability and cross-task generalization performance of molecular property prediction.

The application provides a multi-modal collaborative evolution irony recognition method and device based on granular computing, and relates to the technical field of natural language processing. The method first extracts sample data from multi-modal, completes preprocessing and irony label labeling, forms a multi-modal irony data set aligned across modalities; then uses hierarchical granularity analysis to perform feature clustering in each modal feature space, constructs multi-modal multi-granularity knowledge representation; then filters a number of granularities with the highest dependency from each modality based on a rough set dependency function, generates a multi-modal optimal granularity matrix; then extracts the prototype mode vector of the known irony label and the test mode vector of the to-be-tested sample from the matrix, constructs an irony order parameter that characterizes the similarity between the two; finally, a collaborative neural network model is constructed based on the principle of synergetics, so that each modal order parameter evolves collaboratively under the mechanisms of self-excitation, self-inhibition and lateral inhibition, and the irony labeling mode corresponding to the highest order parameter is output after weight fusion, realizing irony recognition.

This invention provides a visual biomimetic edge detection method based on texture gradient modulation, belonging to the field of image edge detection technology. It focuses on introducing texture gradients to detect significant edges and suppress texture. First, the retina is modeled, and the image is encoded to obtain multiple information channels. Based on this, a novel peripheral modulation mechanism is designed, including unidirectional facilitation, lateral inhibition, and omnidirectional inhibition, to modulate the responses of simple cells in the primary visual cortex. Simultaneously, texture gradients are extracted using texture information and combined with the responses of simple cells to highlight the boundaries of textured regions and weaken the responses at textured edges. Next, endpoint cells in the second-level visual cortex are modeled to further modulate edge responses. Finally, the fourth-level visual cortex is modeled to integrate edge cues from all information channels to obtain the final edge detection result. This invention solves the problems of poorly highlighted textured region boundaries and significant noise in existing edge detection methods.

The application belongs to the technical field of biological information calculation, and discloses a kind of pet feed additive formula computing model optimization system, comprising: data interface obtains raw material parameter and contains environmental stress factor data;Memory stores heterogeneous neural network model, which contains first time series convolution subnet, second cascade processing subnet and global gating distribution unit coupled by lateral inhibition channel and is arranged in parallel;Processor runs model, utilizes global gating distribution unit to dynamically adjust the computing power weight between subnet according to system energy state tensor, and carries out back propagation iteration based on nonlinear constraint penalty term, the special computing architecture of sparse topological mapping and dynamic gating mechanism is constructed, the convergence difficulty and negative transfer problem when general model processes high-dimensional strongly coupled biological data are solved, and the prediction accuracy and safety of formula optimization are improved.

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.

The application belongs to the technical field of image processing, and particularly relates to a soft connection lamination side weld quality detection method based on image processing, which comprises the following steps: constructing a structure tensor according to a gradient component of a target pixel point and obtaining an extension direction through eigenvalue decomposition; performing a morphological closing operation on an original image based on the extension direction to obtain a fitting background value, and calculating a contrast residual in combination with a gray value; searching for a background reference point along a direction perpendicular to the extension direction, and calculating a lateral inhibition weight according to a contrast residual difference between the target pixel point and the background reference point; determining a trajectory cumulative path length according to the contrast residual, calculating a product of the contrast residual of each point on the cumulative path along the extension direction and the corresponding lateral inhibition weight, and obtaining a final defect response value; performing binarization on the defect response graph, and determining the quality of the soft connection lamination according to a connected domain area. The application improves the defect detection accuracy of the soft connection lamination.

This invention discloses an on-orbit intelligent processing method for spaceborne images based on brain-like computing, belonging to the field of on-orbit intelligent processing of spaceborne images. This invention utilizes a spiking neural network to simulate the structure of the human brain: including a feature extraction layer, a pulse coding layer, an STDP learning layer, a lateral inhibition layer, and a decision output layer; and improves model accuracy through on-orbit updates and federated learning. This invention reduces computational energy consumption and improves efficiency. Simultaneously, this invention leverages the synaptic plasticity of neurons to construct a SNN with excellent spatial and temporal resolution.

The invention belongs to the technical field of image processing, and particularly relates to a flexible connection lamination side weld quality detection method based on image processing, and the method comprises the steps: constructing a structure tensor according to the gradient component of a target pixel point, and obtaining an extension direction through eigenvalue decomposition; performing morphological closed operation on the original image based on the extension direction rotation linear structure element to obtain a fitting background value, and calculating a contrast residual error in combination with a gray value; searching a background reference point along a direction perpendicular to the extension direction, and calculating a lateral suppression weight according to a contrast residual difference between the target pixel point and the background reference point; determining the length of a track accumulation path according to the contrast residual error, and calculating the product of the contrast residual error of each point on the accumulation path along the extension direction and the corresponding lateral suppression weight to obtain a final defect response value; and binarizing the defect response diagram, and judging the quality of the flexible connection lamination according to the area of the connected domain. According to the invention, the defect detection accuracy of the flexible connection lamination is improved.

The application relates to an adaptive visual enhancement and target detection method and system based on a center-periphery antagonistic receptive field and a lateral inhibition mechanism, and relates to an adaptive visual enhancement and target detection method and system. In order to solve the problems that the existing method lacks adaptive adjustment, the target detection task is insufficiently coupled, high detection precision and contrast enhancement cannot be ensured at the same time in a complex background, and the biological mechanism is not deeply utilized, the application first simulates a retinal receptive field by using a differential Gaussian filter to realize edge contrast amplification; subsequently, a learnable lateral inhibition layer is added in a deep network to apply inhibition to neighborhood responses to highlight target information. By online adaptive adjustment of the inhibition parameters, the application significantly improves target separability and detection precision in a complex background, and has high robustness and interpretability. The application belongs to the technical field of computer vision.

The invention discloses a pulse neural network system based on mixing of a resonant tunneling diode and a memristor, which belongs to the technical field of neuromorphic calculation and brain-like intelligent hardware and comprises an input module, a pulse shaping and issuing module and a feedback control module. The input module is used for weighting an input pulse signal through a memristor and generating a total post-synaptic potential on a node; the pulse shaping and issuing module performs threshold judgment on the potential and generates standardized pulse output; the feedback control module provides refractory recovery, lateral suppression, and synaptic modulation functions. The system utilizes the memristor to simulate synaptic plasticity and RTD to realize high-speed neuron distribution, has the characteristics of high integration level, low power consumption and fast response, and is suitable for the fields of edge calculation and brain-like intelligent hardware.

The invention provides a multi-modal co-evolution anti-recognition method and device based on granular computing, and relates to the technical field of natural language process.The method comprises the steps that firstly, sample data is extracted from multiple modals, preprocessing and anti-label labeling are completed, and a multi-modal anti-data set with cross-modal alignment is formed; performing feature clustering in each modal feature space by utilizing hierarchical granularity analysis, and constructing a multi-modal multi-granularity knowledge representation; a plurality of granularities with the highest dependency degree are screened from all the modals based on a rough set dependency function, and a multi-modal optimal granularity matrix is generated; extracting a prototype mode vector of a known antitag and a test mode vector of a to-be-tested sample from the matrix, and constructing a heat sequence parameter depicting the similarity of the prototype mode vector and the test mode vector; and finally, constructing a collaborative neural network model based on a synergetics principle, enabling each modal sequence parameter to co-evolve under a self-excitation, self-suppression and side suppression mechanism, and outputting an anti-vital mark mode corresponding to the highest sequence parameter after weight fusion to realize anti-vital mark recognition.