82 results about "Neural decoding" patented technology

A neural classifier model is used to detect cybersecurity vulnerabilities in the source code predicted by a deep learning code generation model having been trained on source code possibly containing security bugs. Upon the classifier model classifying a given source code snippet as likely containing a cybersecurity vulnerability, a proposed repair for the cybersecurity vulnerability is predicted from a neural decoder transformer model having been trained on non-vulnerable source code. The neural decoder transformer model is used to predict source code that repairs the cybersecurity vulnerability given the source code classified with a cybersecurity vulnerability.

The invention discloses a three-dimensional scene reconstruction method based on an intelligent LED street lamp multi-mode sensor. The three-dimensional scene reconstruction method comprises the following steps that RGB images are obtained and preprocessed; the information is input to a visual feature coding module, two-dimensional bounding box information is extracted, and an object segmentation module is guided to output a two-dimensional segmentation mask; obtaining point cloud data, projecting the point cloud data to the standardized RGB image, and screening target points in combination with the two-dimensional segmentation mask; complementing the preliminary segmentation result of the point cloud, mapping the result to a standardized RGB image, and extracting a pixel region; carrying out joint coding, implicit representation and neural decoding processing on the object-level RGB image and the point cloud complete segmentation result; and fusing into an original three-dimensional scene, and completing spatial restoration through point cloud registration, attitude optimization and semantic constraint. The invention provides an efficient three-dimensional scene reconstruction method in combination with a multi-mode sensor of an intelligent LED street lamp, and the method has high precision, real-time performance and dynamic target processing capability.

A speech brain-computer interface neural decoding system based on Chinese language, a method for controlling the speech brain-computer interface neural decoding system, and a method for implementing the speech brain-computer interface neural decoding system are disclosed. The speech brain-computer interface neural decoding system includes an electroencephalography (EEG) data acquisition module, a significance feature screening and verification module, a speech imagery EEG data decoding module, and an understandable speech synthesis module. The speech brain-computer interface neural decoding system integrates EEG data collection, EEG feature extraction, EEG feature screening, EEG signal decoding for reconstructing speech spectrum information and understandable speech synthesis. After obtaining reconstructed spectrogram features, a Pearson correlation analysis may be performed with original spectrogram features, with a correlation generally being ≥80%. The decoding performance of the speech brain-computer interface neural decoding system is superior to traditional decoding models, effectively improving communication between patients with speech disorders and the outside world.

The invention relates to the technical field of medical health data information decoding analysis and processing, and particularly discloses a dynamic pain nerve decoding and intervention system based on a prefrontal lobe time sequence characteristic window, which comprises a signal acquisition module, a dynamic window control module, a phase prediction module, an intervention execution module and an event driving controller. The dynamic window control module intelligently adjusts the duration of a signal analysis window through real-time calculation of the energy ratio of a # imgabs 0 # frequency band to a # imgabs 1 # frequency band, and accurately locks a transient characteristic window in a pain outbreak period; the phase prediction module realizes millisecond-level synchronization of stimulation pulse and a rising edge of action potential of a neural cluster based on a two-factor mapping mechanism of an individualized conduction velocity level and an oscillation frequency; the event-driven controller reduces the power consumption to a wearable device level while maintaining 10 ms-level real-time response through a dynamic resource allocation strategy.

The invention relates to the technical field of video super-resolution reconstruction, in particular to an unmanned aerial vehicle video super-resolution reconstruction method and system based on an implicit diffusion model, and the method comprises the steps: encoding a low-resolution frame into a low-dimensional latent feature through an implicit neural expression auto-encoder; denoising the latent features of the reference frame through an implicit neural U-shaped diffusion network to obtain clean latent features, generating a high-resolution reference image through an implicit neural decoder, and performing optical flow alignment to obtain a time consistency control signal; and finally, inputting the current frame latent features and the control signal into an implicit neural U-shaped diffusion network with a historical sampling correction module, and outputting a super-resolution image of any scale by an implicit neural decoder after iterative denoising. According to the method, in the unmanned aerial vehicle video super-resolution reconstruction task, any-scale reconstruction is realized, the problems of space-time inconsistency and sampling error accumulation are effectively relieved, and the method has better performance in the aspects of perception quality, time sequence consistency and the like.

The invention discloses a progressive migration training decoding method based on GPT, and belongs to the field of optic nerve decoding. According to the method, a GPT-based category and text decoding model is constructed based on functional magnetic resonance signal data recorded when a subject watches a large number of visual perception images, and the method comprises an image encoder used for embedding image feature vectors extracted by adopting a pre-training model into a hidden feature space; prompt containing task information and position codes are embedded into a hidden feature space; the visual information encoder is used for mapping a brain visual region voxel signal to a hidden feature space by using a GRU; the whole-brain information encoder is used for guiding the selection of voxel signals of other regions of the brain by taking a voxel signal of a visual region of the brain as a context, and mapping global information to a hidden feature space; and the language generator fuses all the hidden features and selects the Token with the highest probability to generate the text of the corresponding content. According to the method, the decoding accuracy is improved, and the generalization ability of the model is enhanced.

The invention discloses a brain-controlled speech enhancement method and system based on electroencephalogram signal neural decoding. The method comprises the following steps: synchronously acquiring multi-channel speech signals and EEG signals; compact features related to auditory attention are extracted from the EEG signals through the EEG coding network and serve as auditory cognitive state representation, and an auditory consensus field continuously evolved on the slow time scale is constructed; performing time-frequency modeling on the multi-channel voice signals based on deep learning, and constructing a deep voice time-frequency analysis network; and performing voice time-frequency spectrum gain modulation and neural beam forming weight gain modulation on the signal by using an auditory consensus field, converting the signal back to a time domain, and outputting a finally enhanced target voice signal. According to the invention, by constructing an auditory consensus field driven by the EEG, neural modulation of a frequency spectrum characteristic flow of a deep speech time-frequency analysis network and a space beam forming weight of a neural network beam forming module is realized, so that speech enhancement in a complex acoustic environment and robustness and stability of a system are improved.

The invention relates to an acupuncture brain machine interface system based on brain-like recursive network model decoding. The system comprises an electroencephalogram acquisition module; an electroencephalogram preprocessing module; the needling parameter fitting module is used for accurately fitting needle body motion parameters through a polynomial regression method; the state space model is used for accurately simulating the nonlinear change characteristic of the neuron membrane potential according to the needle body motion parameters output by the acupuncture parameter fitting module; the brain-like recursive network model is used for depicting time dependence of neural activities of a single brain region under the acupuncture action and correlation of neural activities among different brain regions; a Kalman filtering generator; and the neural decoder is used for performing behavior decoding on the optimal internal state output by the brain-like recursive network model updated by the Kalman filtering generator. The system integrates electroencephalogram signal acquisition, acupuncture parameter acquisition, brain nerve activity modeling and acupuncture manipulation decoding, can perform real-time online operation, and is high in time resolution and high in response speed.

The invention provides a clinical motion control device, method and system, electronic equipment and a medium, and the device comprises an ultra-micro array implantation probe which is configured to be implanted into a brain motor cortex of a patient and collect a neuroelectric signal of the brain motor cortex; the integrated data processing module is electrically connected with the ultra-micro array implanted probe and is configured to convert the neuroelectric signals into digital signals and perform neural decoding on the digital signals to obtain motion control signals containing the motion intention of the patient; and the data transmission module is in communication connection with the integrated data processing module and is configured to transmit the motion control signal to external rehabilitation equipment through a wired interface or a wireless communication mode, so that the external rehabilitation equipment drives the limbs of the patient to execute corresponding rehabilitation motions. According to the invention, large-scale, high-precision and synchronous monitoring can be carried out on brain motor cortex neural activities, the signal acquisition quality is improved through an integrated structure design, and the signal decoding precision is improved at the same time.

The invention provides a neural speech compression method and system, and the method comprises the steps: obtaining original speech data, and carrying out the encoding of the original speech data through an encoder, and obtaining a hidden vector; inputting the hidden vector into a layered residual vector quantization module, and quantizing the hidden vector into a discrete code based on a multi-scale parallel token stream; inputting the discretized code into a decoder for decoding so as to reconstruct a language signal and obtain a target language signal; wherein the encoder is configured to introduce a cross-modal distillation technology to express high-level semantics extracted from a teacher network model when the original voice data is encoded; the decoder is configured as a neural decoder that performs adversarial training enhancement with the multi-scale discriminator. According to the method, the potential structure of the signal can be learned by utilizing deep learning, breakthrough progress is achieved in a low-bit-rate scene, the modeling capability and the compression quality are improved, and the method can be widely applied to scenes such as voice communication, audio storage and edge intelligent equipment.

The invention discloses an extreme image coding and decoding method based on a diffusion model. The method comprises the following steps: S100, a compression and pre-decoding module at a coding end extracts, compresses and transmits image information by using a neural coding network; s200, a compression and pre-decoding module of the decoding end preliminarily decodes the received image information into a content variable aligned with the potential diffusion space by using a neural decoding network; and S300, a denoising module uses the content variable as a control condition and a text graph diffusion model as prior information, and reconstructs an image from random noise through step-by-step denoising. The method can utilize the strong generation capability of the pre-trained text image diffusion model to realize perception-friendly reconstruction consistent with the original image at an extremely low bit rate, has the characteristics of ultrahigh compression ratio and good image reconstruction effect, can carry out image transmission under an extremely low bandwidth condition, and has a wide application prospect. The method can be widely applied to the fields of smart phone satellite communication, short-wave communication, unmanned equipment remote operation and the like.

An automated system for resolving program merges uses neural transformers with attention. In one aspect, a neural encoder transformer model is trained from developer-resolved merge conflicts to learn to predict a resolution strategy that aids a developer in constructing a merged program. In a second aspect, a neural decoder transformer model is trained on the syntax and semantics of different source code programming languages to predict a merge resolution consisting of interleaved lines of source code from programs A, B, or O, where programs A and B contain changes to code base O.

The invention provides an EEG characterization method based on wavelet neural quantization training and semantic alignment, and the training method comprises the steps: S1, obtaining training data which comprises a plurality of EEG signal samples, and each sample comprises a plurality of EEG sub-signals; s2, using the training data to train a wavelet quantization nerve marker and a nerve codebook thereof: using a vector quantization encoder of the wavelet quantization nerve marker to carry out discrete wavelet transform on input samples, generating a feature block obtained by discrete wavelet transform and corresponding to each sub-signal of each sample, and extracting a potential representation of each feature block; acquiring a closest discrete vector matched from a plurality of discrete vectors in the neural codebook according to the potential representation, and reconstructing the EEG sub-signal corresponding to the feature block according to the discrete vector matched with each feature block by using a neural decoder based on inverse wavelet transform; and training and updating the parameters of the nerve marker and discrete vectors of the nerve codebook according to the first total loss function.

The invention discloses a multivariable neural decoding method based on OPM-MEG and RSA, and belongs to the technical field of biomedical engineering. The method comprises the following steps: acquiring resting state structural MRI data and an OPM-MEG scanning image for joint registration; playing auditory stimulation, completing OPM-MEG data acquisition, and preprocessing the data to obtain segmented data; averaging different test times of the same auditory stimulation on a segment data level, constructing a neural response mode vector, and calculating the Pearson correlation of all possible pairwise stimulation pairs at time points to obtain a representation similarity matrix and a neural response mode similarity average result; determining time points with differences one by one; a sequence of source activation times for each subject is estimated, and activation regions are determined using T-tests. According to the method, the decoding representation process of the brain can be disclosed by calculating the distance of the neural reaction mode in the representation space, so that the response mode of the brain under a specific task or condition can be comprehensively understood.

Systems and methods for high-bandwidth, minimally invasive brain-computer interfaces (BCIs) are disclosed. The BCIs are configured for deployment and operation in conjunction with a comprehensive interventional electrophysiology procedural suite. Three primary methods of minimally invasive electrode array delivery are disclosed: (1) cortical surface delivery, (2) ventricular delivery, and (3) endovascular delivery. Additionally, systems and methods for interacting with such high-bandwidth electrode arrays are discussed, including real-time imaging, signal processing, and neural decoding. Systems and methods for architectures for accelerating the underlying computational processes (such as graphics processing units or tensor processing units) are also discussed. Multiple applications of BCIs are discussed, with emphasis on restoration, rehabilitation, and augmentation of neurologic function.

The present invention discloses a multivariate neural decoding method based on OPM-MEG and RSA, which belongs to the field of biomedical engineering technology. The method comprises: obtaining resting-state structural MRI data and OPM-MEG scan images for joint registration; playing auditory stimulation and completing OPM-MEG data acquisition, preprocessing the data, and obtaining segmented data; averaging different trials of the same auditory stimulation at the segmented data level and constructing a neural response pattern vector, calculating the Pearson correlation of all possible pairwise stimulus pairs at each time point, obtaining a representation similarity matrix and an average result of the neural response pattern similarity; determining the time points with differences at each time point; estimating the source activation time series of each subject, and determining the activation area using a T-test. The present invention can reveal the decoding representation process of the brain by calculating the distance of the neural response pattern in the representation space, thereby more comprehensively understanding the response pattern of the brain under specific tasks or conditions.

The invention particularly relates to a neural decoding method based on multi-scale electroencephalogram signals and heterogeneous visual stimulation alignment, which comprises the following steps: firstly, performing frequency domain transformation, processing and inverse mapping on original EEG signals through an FADE module to effectively integrate frequency specific information; then, the signals are input into an ACC module, and the module achieves self-adaptive multi-channel relation modeling based on a functional relation instead of a fixed anatomical position through a dynamic clustering center and a cross attention mechanism; then, the encoded EEG features and CLIP visual features are subjected to alignment training under a joint loss function, and the loss function optimizes the semantic fidelity and discrimination of the features at the same time; finally, the EVA frame obtained through training can be used for heterogeneous visual stimulation decoding tasks such as zero-sample image retrieval, video classification and high-quality image reconstruction. According to the method, the capture capability of the EEG features on visual semantic information and the generalization of the model can be effectively improved, and the challenge of multi-scale and multi-modal neural decoding is solved to a certain extent.

The application discloses a kind of three-dimensional scene reconstruction methods based on intelligent LED street lamp multimodal sensor, comprising the following steps: obtaining RGB image and pre-processing;Input to visual feature coding module, extract two-dimensional boundary box information, guide object segmentation module to output two-dimensional segmentation mask;Obtain point cloud data, project to standardized RGB image, in combination with two-dimensional segmentation mask screening target point;Complete point cloud preliminary segmentation result, map to standardized RGB image, extract pixel area;The RGB image and point cloud complete segmentation result of object level are jointly coded, implicit representation and neural decoding processing;Fusion into original three-dimensional scene, complete space restoration by point cloud registration, pose optimization and semantic constraint.The application combines the multimodal sensor of intelligent LED street lamp, and proposes a kind of efficient three-dimensional scene reconstruction method, with high precision, real-time and dynamic target processing capacity.

The present invention discloses a wearable medical device for treating peripheral pain, including a stimulation strategy unit, an electrode array unit, a pulse excitation unit, an impedance measurement unit, an adaptive threshold anchoring unit, an interference intensity detection unit, and a solution unit. Through lightweight wearable technology and non-invasive deep neural regulation technology, specific intervention and regulation are performed on peripheral nerves to achieve effective relief of peripheral pain. The present invention combines non-invasive wearable technology, spatial interference electrical technology, spatial electrode arrays, and neural decoding stimulation strategies to achieve non-invasive treatment of peripheral pain.

Systems and methods for high-bandwidth, minimally invasive brain-computer interfaces (BCIs) are disclosed. The BCIs are configured for deployment and operation in conjunction with a comprehensive interventional electrophysiology procedural suite. Three primary methods of minimally invasive electrode array delivery are disclosed: (1) cortical surface delivery, (2) ventricular delivery, and (3) endovascular delivery. Additionally, systems and methods for interacting with such high-bandwidth electrode arrays are discussed, including real-time imaging, signal processing, and neural decoding. Systems and methods for architectures for accelerating the underlying computational processes (such as graphics processing units or tensor processing units) are also discussed. Multiple applications of BCIs are discussed, with emphasis on restoration, rehabilitation, and augmentation of neurologic function.

The invention discloses a neural rehabilitation method based on neural decoding and a three-dimensional generative technology, and relates to the technical field of neural rehabilitation, and the method mainly comprises the steps: collecting a multi-channel electromyographic signal of a patient, and outputting a predicted gesture category label according to the time domain feature and the frequency domain feature extracted from the multi-channel electromyographic signal; adopting HaMeR to obtain a hand shape parameter of the patient, and combining the hand shape parameter with the posture parameter corresponding to the gesture category label and the posture parameter of the example gesture to reconstruct a three-dimensional patient gesture and a three-dimensional example gesture; taking two adjacent different three-dimensional patient gestures as an initial gesture and a target gesture, and generating a smooth motion track between the initial gesture and the target gesture through a diffusion model; and synchronously displaying the visual smooth motion trail, the motion video of the three-dimensional example gesture and a comparison thermodynamic diagram of the myoelectricity root-mean-square value of the patient and the healthy population data through an interactive interface. A precise electromechanical structure of a traditional exoskeleton is replaced with the virtual interaction interface, and a technical basis is provided for general rehabilitation medical treatment.

Systems and methods for high-bandwidth, minimally invasive brain-computer interfaces (BCIs) are disclosed. The BCIs are configured for deployment and operation in conjunction with a comprehensive interventional electrophysiology procedural suite. Three primary methods of minimally invasive electrode array delivery are disclosed: (1) cortical surface delivery, (2) ventricular delivery, and (3) endovascular delivery. Additionally, systems and methods for interacting with such high-bandwidth electrode arrays are discussed, including real-time imaging, signal processing, and neural decoding. Systems and methods for architectures for accelerating the underlying computational processes (such as graphics processing units or tensor processing units) are also discussed. Multiple applications of BCIs are discussed, with emphasis on restoration, rehabilitation, and augmentation of neurologic function.

A part of a data stream includes a plurality of data units respectively associated with a plurality of images and together presenting different images of the plurality of images. The data stream part further includes differential encoding data differentially encoding an artificial neural decoding network relatively to a reference artificial neural decoding network. A method for decoding this data stream part includes the following steps: determining the artificial neural decoding network by decoding the differential coding data; and decoding at least one data unit of the plurality of data units by the determined artificial neural decoding network. Another decoding method, decoding devices, a computer program and associated data streams are also described.

The invention discloses an fMRI video nerve decoding method based on visual perception and semantic consistency. The method comprises the following steps: acquiring an fMRI signal and corresponding video data; an fMRI input vector is constructed; extracting a color map and a space depth map from the video data, and embedding frame semantics and text description semantics; the fMRI input vector is used as a feature, the color map and the space depth map are used as labels, a visual perception path is trained based on a double-stage trained double-tower mask auto-encoder, and prediction of the color map and prediction of the space depth map are obtained; the fMRI input vector is used as a feature, frame semantic embedding and text description semantic embedding are used as tags, a semantic path is trained, and prediction of fMRI semantic embedding is obtained; and performing multi-condition guidance on the video diffusion model based on prediction of the color map, the space depth map and fMRI semantic embedding to obtain decoding output. According to the method, fMRI video nerve decoding with high visual perception consistency and semantic consistency can be realized.

A method for encoding a video signal includes receiving encoding instructions for encoding the video signal and encoding the video signal using a training loop for optimizing a metric of the encoded video signal, the training loop including a loss function and one or more of a neural encoder and a neural decoder, and the training loop is adapted based on the encoding instructions.

The invention discloses a neural decoding method based on three-space decoupling and multi-mode mutual information. The method comprises the following steps: respectively coding multi-modal data to obtain a private representation, a shared representation and a collaborative representation, and correspondingly converting the private representation, the shared representation and the collaborative representation into distribution parameters of a private space, a shared space and a collaborative space; obtaining a potential representation based on the private representation, the shared representation and the collaborative representation, decoding the potential representation into reconstruction data, and obtaining a distribution parameter of the reconstruction data; constraining private space mutual information, constraining shared space mutual information and constraining collaborative space mutual information; constructing a total loss function to train the private space, the shared space and the collaborative space; and performing visual category neural decoding based on a dynamic feature enhanced attention fusion method. According to the method, the accuracy of fMRI-image cross-modal representation learning and visual category neural decoding can be improved.

The application discloses an eye movement behavior neural decoding method, device and equipment, comprising: synchronously collecting behavior monitoring video data and multi-channel brain activity signals of a subject in a natural scene; performing spatiotemporal feature analysis on the behavior monitoring video data to generate an eye movement feature sequence representing discrete eye movement events; inputting the multi-channel brain activity signals into a preset whole brain generation model to generate a brain network atlas representation including a whole brain stereotactic electroencephalogram signal and an initial brain network; inputting the eye movement feature sequence, the whole brain stereotactic electroencephalogram signal and the initial brain network into a pre-trained spatiotemporal mapping decoding model to obtain spatiotemporal joint features; and inputting the spatiotemporal joint features into a classifier to output an eye movement behavior classification result of the subject. The accuracy and robustness of eye movement behavior decoding can be improved.

The present invention relates to the technical field of neural decoding, and particularly relates to a lightweight motor imagery classification method and device guided by frequency prior features. The method includes: extracting the time-domain features of high-dimensional EEG signals; performing weighted integration on the EEG signals collected by different electrodes through a spatial convolutional layer; performing downsampling through average pooling to compress the high-frequency information of the weighted integrated EEG signals; inputting the compressed EEG signals into a residual network to process the high-frequency information of the compressed EEG signals and further optimize the extraction of signal features; designing causal convolution and dilated convolution to effectively capture the temporal features of EEG signals and ensure that the network structure has good computational efficiency and interpretability. This method provides an efficient and interpretable solution in EEG signal analysis and can significantly improve the performance of the model.

The invention provides a post-injury rehabilitation tracking system and method based on high-density surface electromyographic signals, and belongs to the technical field of sports medicine and rehabilitation engineering, and the system comprises a synchronous collection module which is used for synchronously collecting the high-density electromyographic signals and joint torque signals; the signal preprocessing module is used for preprocessing the electromyographic signals; the neural decoding module is used for constructing a space-time expansion vector, updating signal statistical characteristics and separating to obtain a discharge sequence of the motion unit; the function evaluation module is used for calculating and optimizing a nerve density index; the rehabilitation decision output module is used for calculating indexes of the indexes and generating personalized training suggestions according to whether the indexes reach the standard or not; and the parameter adjusting module is used for receiving the personalized training suggestions and continuously optimizing the suggestions based on the execution feedback information. And the efficiency, the accuracy and the individuation degree of clinical rehabilitation evaluation and training control are remarkably improved.