915 results about "Encoder decoder" patented technology

Disclosed in the present invention is a non-autoregressive Transformer-based modeling method for a 4-level pulse amplitude modulation high-speed transmitter. The method involves establishing a deep learning model having an encoder-decoder architecture to predict the behavior of a 4-level pulse amplitude modulation transmitter. An encoder processes unordered non-sequential inputs, including an input signal parameter and link parameters, to generate a context vector and then transmit same to a decoder. The decoder uses both the context vector generated by the encoder and a transmitter output signal sequence to generate a categorical probability distribution for each point in the sequence one by one. The model is trained using a random masking strategy, and inference is performed by means of non-autoregressive decoding and filtering, so that the model can perform parallel prediction on an output sequence, and perform a filtering process to predict an output signal. Compared to traditional simulation methods, the present invention achieves a significant acceleration effect, particularly when processing multi-link systems.

The invention relates to a medical image segmentation method and system based on residual Mama and multi-scale boundary enhancement. The method comprises the following steps: acquiring and preprocessing a medical image; inputting the image into a segmentation model based on an encoder-decoder architecture; the encoder synchronously extracts local texture features and models long-range spatial dependence through residual error convolution blocks and residual error Mama blocks which are alternately connected; fusing and enhancing the jump connection features between the encoder and the decoder through a boundary enhancement module to optimize boundary characterization; integrating a multi-scale gating attention module in a decoding path, and adaptively selecting and fusing multi-scale context features; and finally outputting the high-precision segmentation mask. The method effectively solves the problems that in the prior art, long-range dependence and local details are difficult to consider, the multi-scale feature fusion capability is insufficient, boundary segmentation is fuzzy and the like, and the segmentation accuracy, the boundary continuity and the clinical practicability are remarkably improved.

The invention relates to the technical field of photovoltaic prediction, in particular to a distributed photovoltaic power prediction method and system based on high-dimensional gridding numerical weather forecast, and the method comprises the steps: carrying out the standardization of the numerical weather forecast data and photovoltaic power historical data of a target region, and achieving the time-space alignment based on a preset grid, generating a gridding data set; utilizing convolution processing to extract local space features, and converting and fusing the local space features into a feature sequence containing space and historical time sequence information at the same time; modeling is carried out through an encoder-decoder architecture, an encoder excavates historical power dependence, and a decoder dynamically couples future meteorological characteristics with historical power through an attention mechanism and outputs a grid-level predicted value; aggregating to obtain a system total power prediction result; by establishing a unified space-time grid, refined alignment of data is realized, cross-space-time dynamic fusion is performed in combination with convolution and an attention mechanism, and prediction precision and stability can be kept in complex weather.

The invention discloses a remote sensing image target detection method based on RT-DETR, and the method comprises the following steps: 1, obtaining a remote sensing image data set, and completing the preprocessing; 2, inputting a backbone network, and extracting a multi-scale feature map; step 3, inputting a DSF module to realize multi-scale feature adaptive fusion; step 4, inputting an MSFE module, enhancing features and modeling long-range dependence; 5, embedding a G-HCO module in a corresponding stage of the backbone network, and optimizing feature expression; step 6, inputting an RT-DETR encoder-decoder, and outputting a target category and a target position; and 7, filtering the low-confidence prediction frame to obtain a final detection result. According to the remote sensing image target detection method based on the RT-DETR, through multi-scale feature adaptive fusion of the DSF module, frequency domain-spatial domain dual-domain feature enhancement of the MSFE module and global feature optimization of the G-HCO module, the method can also be expanded to multi-task scenes such as video target detection and segmentation-detection combination, and the generalization ability and the application range are better.

The invention discloses a complex network disintegration method based on evolution deep reinforcement learning. According to the method, an encoder-decoder model fusing a graph convolutional neural network and a deep Q network is constructed, and is used for efficiently extracting importance features of nodes in a complex network and realizing dynamic decision-making of a node disassembling sequence according to the importance features. In order to optimize model parameters and improve search capability, an evolutionary algorithm is introduced to perform global exploration on the model parameters, and the problem that a directional optimization strategy is easy to fall into local optimum is avoided. Meanwhile, deep mining is carried out on an evolution result in combination with a reinforcement learning strategy, the overall optimization process is accelerated, and advantage complementation of parameter evolution and strategy learning is achieved. Experimental results show that the method significantly improves the efficiency and precision of network disassembly while maintaining the robustness of the model, and has good practical value and wide application prospects.

The invention discloses a manifold perception-based large model navigation method guided by a historical topological graph, and relates to a computer vision technology. The method aims at solving the challenges that in the navigation process, long-distance reasoning experience is insufficient, instruction fragments and dynamic visual observation are difficult to align, and large model reasoning is prone to illusion interference. Firstly, a large model based on an encoder-decoder structure is used for supplementing historical information coding for a visual observation sequence, and therefore global topological information guidance is provided for long-distance reasoning. And secondly, in order to effectively solve the problem that large model reasoning is subjected to illusion interference, significant space-time differences in a visual observation sequence are mined by using a multi-curvature manifold, so that the large model can accurately describe the current environment and make a decision according to a visual reference object in thinking. Besides, in order to strengthen the perception capability of the large model to the space structure and establish a graph self-attention mechanism, the node distance embedded in the constructed historical topological graph is combined with the visual similarity so as to model the space relationship between the nodes.

An encoder includes circuitry and memory coupled to the circuitry. In operation, the circuitry encodes subpicture information in which a horizontal position and a vertical position of a region of a subpicture are represented in a unit of a coding tree unit (CTU). The subpicture is a rectangular region in a picture. The horizontal position is represented by a first position of a first CTU in the subpicture, the first position being relative to a left end of the picture. The vertical position is represented by a second position of a second CTU in the subpicture, the second position being relative to a top end of the picture.

Picture coding and decoding methods, an encoder, a decoder and a storage medium are provided. The encoder, before performing coding processing according to a matrix-based intra prediction (MIP) mode, sets initial right shift parameters corresponding to different sizes and different MIP mode numbers as an uniform offset parameter, the offset parameter indicating a number of right shifting bits of a predicted value, and when performing coding processing according to the MIP mode, performs coding processing according to the offset parameter. The decoder, before performing decoding processing according to an MIP mode, sets initial right shift parameters corresponding to different sizes and different MIP mode numbers as an uniform offset parameter, the offset parameter indicating a number of right shifting bits of a predicted value, and when performing decoding processing according to the MIP mode, performs decoding processing according to the offset parameter.

The invention discloses an SAR image despeckle method and device, and relates to the technical field of image processing. The method comprises the following steps: constructing an SAR image despeckle model based on an encoder-decoder structure; acquiring an SAR observation image polluted by noise; extracting shallow layer features of the SAR observation image; separating the shallow features into high-frequency features and low-frequency features; aggregating the spatial context information of the high-frequency features and encoding the spatial context information into a query matrix, a key matrix and a value matrix; calculating high-frequency output characteristics according to cross covariance attention among the matrixes; extracting noise distribution features in the low-frequency features and splicing the noise distribution features with the low-frequency features to obtain low-frequency output features; learning complementarity weight coefficients of the high-frequency output features and the low-frequency output features, and fusing the high-frequency output features and the low-frequency output features to obtain deep features; reconstructing the deep features output by the last layer of the decoder, and outputting a residual image; and connecting the residual image with the SAR observation image residual to obtain an SAR freckle-removed image.

The invention relates to a transient electromagnetic and seismic wave multi-mode joint inversion imaging method based on physical information Transform, and belongs to the crossing field of geophysical exploration and artificial intelligence. Comprising the following steps: carrying out anomaly detection, interpolation, filtering and normalization preprocessing on transient electromagnetic and seismic wave original data; extracting features representing electrical property, elasticity and cross physical significance; serializing the spatial data through a gridding and alternating fusion strategy, and constructing an enhanced code fusing absolute and relative positions and physical attributes; a physically constrained encoder-decoder architecture is designed to carry out multi-scale forward modeling-inversion; and quantizing the uncertainty of an inversion result by adopting a Bayesian Monte Carlo method, and generating a confidence map. According to the method, through physical rule driving and multi-modal depth complementary fusion, the fine recognition capability and interpretation reliability of hidden disaster-causing geologic bodies such as underground goaf and collapse columns are effectively improved while the physical consistency of data is kept.

The invention discloses a CBCT tooth segmentation method and system based on an anatomical perception cascade network, and the method comprises the steps: a first stage, carrying out the simplified dichotomy segmentation based on an original CBCT image and a coarse segmentation network taking 3D U-Net as a trunk, outputting maxillary and mandibular tooth probability graphs, and taking the maxillary and mandibular tooth probability graphs as prior information to guide the generation of an SDM; in the second stage, the original CBCT image and the calibrated maxillary tooth probability graph and the calibrated mandibular tooth probability graph are spliced together, a formed multi-channel input tensor is input into a fine segmentation network, the fine segmentation network takes Residual U-Net as a trunk, and an improved AGBR module and an improved SDMAA module are integrated in an encoder-decoder architecture of the fine segmentation network; and the decoder fuses all refined and re-calibrated feature maps, and upsamples and reconstructs 42 types of instance segmentation results with correct topology and clear boundaries. According to the method, the problem that in the prior art, when the inherent and local boundary fuzzy defect in CBCT is overcome, an effective pertinence mechanism is lacked is solved, and precise and robust 42-class instance segmentation can be achieved.

The invention relates to the field of image deblurring, and discloses a multi-scale image deblurring method based on a potential space condition diffusion model, comprising the following steps: constructing a multi-scale image deblurring network which comprises a condition diffusion model and a sliding window attention module, the conditional diffusion model is used for generating a multi-scale prior feature from the fuzzy condition vector in a potential space; the sliding window attention module is a U-shaped network based on an encoder-decoder and is used for executing image deblurring feature regression guided by multi-scale prior features; training the network by adopting a two-stage strategy comprising pre-training and post-training; and inputting a blurred image to be processed into the trained multi-scale image deblurring network, and outputting a final deblurred image. According to the method disclosed by the invention, the common problems of excessive smoothness and artifacts in the deblurring process can be effectively inhibited while the calculation efficiency is ensured, and the recovery precision of texture details and edge structures is improved.

The invention discloses an insulator defect identification method for an UHV power transmission line unmanned aerial vehicle inspection image, and belongs to the technical field of power transmission line inspection. The method comprises the following steps: transmitting a visible light image and an infrared image of an insulator shot by an unmanned aerial vehicle in the inspection process to an edge computing platform on the unmanned aerial vehicle; constructing a double-branch feature extraction network architecture to extract an insulator visible light image feature map and an insulator infrared image feature map; constructing a space-time characteristic interaction module to output a deformation visible light characteristic pattern and an infrared time sequence characteristic pattern; performing feature fusion on the deformed visible light feature map and the infrared time sequence feature map to obtain a fused feature map; inputting the fused feature map into a network model based on an encoder-decoder architecture to output a high-resolution feature map; and performing defect identification based on the high-resolution feature map. According to the invention, autonomous identification of insulator defects in the unmanned aerial vehicle inspection process is realized, and the inspection efficiency and accuracy are improved.

Picture coding and decoding methods, an encoder, a decoder and a storage medium are provided. The encoder, before performing coding processing according to a matrix-based intra prediction (MIP) mode, sets initial right shift parameters corresponding to different sizes and different MIP mode numbers as an uniform offset parameter, the offset parameter indicating a number of right shifting bits of a predicted value, and when performing coding processing according to the MIP mode, performs coding processing according to the offset parameter. The decoder, before performing decoding processing according to an MIP mode, sets initial right shift parameters corresponding to different sizes and different MIP mode numbers as an uniform offset parameter, the offset parameter indicating a number of right shifting bits of a predicted value, and when performing decoding processing according to the MIP mode, performs decoding processing according to the offset parameter.

The invention discloses an unmanned aerial vehicle group elastic path planning method and system for resisting an intelligent jammer, which can ensure the accurate synchronization of an interference strategy while realizing the efficient compression and transmission of information and reducing the communication bandwidth demand by an interference party by means of an encoder-decoder communication mechanism, and cooperate with a preferential experience playback technology to realize the flexible path planning of the unmanned aerial vehicle group. Learning is rapidly carried out from past experiences, convergence of an interference strategy is accelerated, efficient cooperative interference is achieved, and a strong threat is formed for a data collection task of the unmanned aerial vehicle group; at the defense party, an attention reviewer module is constructed based on an advanced attention mechanism, information having key influence on unmanned aerial vehicle decision making is accurately focused, the attention weight between intelligent agents is calculated, the weight is reasonably distributed for a state-action joint vector, and a more targeted decision is made in a complex environment; a double-commentator module is introduced, a value function is independently estimated by using two commentator networks, and the reliability and the stability of the algorithm are improved through smooth function fusion output and stable value function estimation.

The invention relates to the technical field of load prediction, and provides a power load prediction method and system based on time sequence decomposition and an attention mechanism, and the method comprises the steps: carrying out the adaptive time sequence decomposition of an obtained original load sequence, calculating the sample entropy of each decomposed component, and carrying out the clustering; constructing a group of encoder and decoder networks for each piece of clustered data, performing parallel encoding to extract features, performing serial decoding reconstruction on the features from low frequency to high frequency, and outputting prediction data from low frequency to high frequency step by step; the weight is initialized based on the sample entropy, the trained weight is obtained through optimization in the encoder and decoder network training process, and the predicted value of the power load is obtained through weighted fusion. According to the method, adaptive time sequence decomposition, a weight mechanism guided by sample entropy and an attention-enhanced encoder-decoder structure are introduced, so that multi-component collaborative modeling and cross-scale dynamic prediction are realized, and the prediction accuracy and stability in complex load data and small sample scenes are effectively improved.

The invention discloses a medical image optimization method and system based on vascular branch selective blurring. The method comprises the following steps: carrying out preprocessing and blood vessel enhancement on a three-dimensional angiography DICOM image; performing precise blood vessel segmentation by using an encoder-decoder network comprising a wide activation and residual cavity space pyramid module; identifying an interference branch which shields the target blood vessel structure at a specific working angle through blood vessel topology analysis; selective blurring repair is performed on the interference branches based on a Poisson equation and a bidirectional convolution LSTM to generate an optimized image which is visually unobstructed and keeps topological continuity. And a safety mechanism of virtual-real combined display and multi-angle plan planning is introduced, so that the reliability of surgical navigation is ensured. And finally, integrating the optimized 3D blood vessel model to a radiography system supporting real-time synchronization, and using the 3D blood vessel model as a road map. The clinical problem that the optimal working angle cannot be used due to blood vessel shielding can be effectively solved, and the precision and safety of an endovascular interventional operation are remarkably improved.

Disclosed in embodiments of the present application are an encoding method, a decoding method, a code stream, an encoder, a decoder, and a storage medium. The decoding method includes: decoding a relevant syntax element of a current block; according to the relevant syntax element, determining to use an intra-template matching prediction (TMP) (IntraTMP) merged intra prediction mode to perform prediction on the current block; determining a matching block of the current block on the basis of a TMP mode, and determining a first predicted block of the current block according to the matching block; determining a second predicted block of the current block on the basis of a non-template matching intra prediction mode; and merging the first predicted block and the second predicted block to determine a final predicted block of the current block.

Picture coding and decoding methods, an encoder, a decoder and a storage medium are provided. The encoder, before performing coding processing according to a matrix-based intra prediction (MIP) mode, sets initial right shift parameters corresponding to different sizes and different MIP mode numbers as an uniform offset parameter, the offset parameter indicating a number of right shifting bits of a predicted value, and when performing coding processing according to the MIP mode, performs coding processing according to the offset parameter. The decoder, before performing decoding processing according to an MIP mode, sets initial right shift parameters corresponding to different sizes and different MIP mode numbers as an uniform offset parameter, the offset parameter indicating a number of right shifting bits of a predicted value, and when performing decoding processing according to the MIP mode, performs decoding processing according to the offset parameter.

The invention discloses a structure perception mask reconstruction learning system and method for OCT image segmentation, and belongs to the technical field of image processing. According to the system, in a pre-training stage, an encoder fused with ViT-Base and a multi-expert self-adaptive reconstruction module is utilized to perform mask reconstruction on a label-free OCT image. In the pre-training stage, two-stage training is performed through a random mask strategy and an attention-guided dynamic mask strategy, so that an encoder-decoder can be transited from random learning to focusing on a structural key area, and optimization is performed by combining a structural perception loss function, thereby forcing a model to learn fine boundary and texture information of a retina. After training is completed, an encoder which learns rich structure knowledge serves as a backbone network to be migrated and is combined with a segmentation decoder of a downstream task, fine adjustment is carried out on a small amount of annotation data, and the accuracy of downstream segmentation is remarkably improved.

The invention relates to a medical image segmentation system and method based on context and frequency guidance and a storage medium, and the system employs an encoder-decoder architecture, and integrates three core modules: a visual state space module which captures global context information with linear complexity based on a selective state space model; the frequency-guided representation module is used for explicitly separating and enhancing structure and boundary characteristics through frequency domain transformation and complex weight modulation; and the multi-scale adaptive context aggregation module is used for integrating multi-scale semantics through parallel multi-branch convolution and dual Top-K sparse attention and focusing a key region. According to the method, the technical problems that global modeling and boundary details are difficult to consider, the calculation complexity is high and the adaptability to multi-scale targets is poor in the prior art are solved, the segmentation precision and the boundary description accuracy are remarkably improved in various medical image segmentation tasks such as heart MRI, polyp, skin lesion and pathological sections, and the medical image segmentation efficiency is improved. And meanwhile, the calculation efficiency is ensured.

The invention discloses a GRU-based vehicle-mounted CAN intrusion detection method and system, a storage medium and a computer system. According to the method, an automatic encoder (AE) is introduced to deepen the understanding of a model on input sequence characteristics, a sliding window is used for selecting batch CAN data to be preprocessed to obtain 13-dimensional time sequence data, and a scalar value within the range of [0, 1] is obtained through processing of the encoder, a GRU, a decoder, a full connection layer and a sigmoid activation function and used for classification of abnormal data. The Conv1D is used as a hidden layer, and compared with two-dimensional convolution, the one-dimensional convolution parameter quantity is smaller, and the calculation is simpler and more convenient. An attack message and a normal message can be completely distinguished, the precision and the accuracy rate reach 100%, and the precision in Fuzz detection is 0.9983; compared with the prior art, the method has high accuracy and reliability in the aspect of intrusion behavior detection, can effectively identify most intrusion events, and can keep a relatively low overall error rate, so that good balance between safety and availability is realized.

The invention provides a deep learning driven adaptive inspection method and system for tunnel crack recognition, and belongs to the technical field of tunnel crack recognition, and the method comprises the steps: collecting an image in a tunnel in real time, and carrying out the preprocessing of the collected image; inputting the preprocessed image into a deep learning crack recognition model to obtain a recognition result; wherein the deep learning crack identification model comprises an encoder, a decoder and an output end; the encoder encodes the input preprocessed image and outputs four feature layers with different resolutions, and the details, the trend, the overall form and the global context of the crack are reserved respectively; the decoder performs up-sampling and fusion on the four feature layers, recovers high-resolution features, and generates a crack segmentation mask; calculating an image definition score, identification uncertainty and a crack continuity index; and controlling the speed of inspecting the image in the tunnel based on the calculated image definition score, the recognition uncertainty and the crack continuity index self-adaptive speed.

The invention discloses a multi-modal feature adaptive fusion and compression semantic communication and classification method and device, and the method comprises the steps: carrying out the multi-modal feature extraction of a collected multi-modal signal, and obtaining a multi-modal feature; carrying out adaptive fusion on the multi-modal features to obtain fused features; and entropy coding compression is performed on the fused features to obtain fused compressed features, and the fused compressed features are converted into binary code streams through an encoder-decoder mechanism for realizing semantic communication with a task module. The method has the advantages of being high in fusion efficiency, friendly in compression, low in calculation overhead, small in communication cost and bandwidth occupation and the like, and the dual requirements for intelligent sensing and task execution capacity under the condition that the bandwidth is limited or the computing power is limited are met.

The invention relates to the field of remote sensing image semantic segmentation of computer vision, in particular to a post-disaster road segmentation method and device based on edge driving and layered global reconstruction. Comprising the following steps: acquiring a remote sensing image of a post-disaster scene, and carrying out feature encoding and decoding on the remote sensing image through an encoder-decoder structure; in the encoding stage, edge features of an image are extracted, edge prior information representing the boundary position and strength of a target is generated, and weighting constraint is carried out on shallow encoding features according to the edge prior information, so that edge constraint features are obtained. And carrying out multi-scale extraction and fusion on the features to form fusion features containing different spatial scale information. In the decoding stage, global semantic description information is generated based on the fusion features, and hierarchical guide reconstruction is performed on the fusion features by using the information to obtain reconstructed features. And generating a segmentation result of the post-disaster road according to the reconstruction features and outputting the segmentation result. The pixel-level segmentation effect of clear boundary, continuous structure and high robustness of the road area in a complex post-disaster scene is realized.

Systems and methods for identifying anomalous activities in a cloud computing environment are provided. According to one embodiment, a customer's infrastructure may be fortified by leveraging deep learning technology (e.g., an encoder-decoder machine-learning (ML) model) to predict events in the cloud environment. During a training phase, the ML model may be trained to make a prediction regarding a next event based on a predetermined or configurable length of a sequence of contextual events. For example, historical events (e.g., cloud application programming interface (API) events logged to a cloud activity trace) observed within the customer's cloud infrastructure over the course of a particular date range may be split into appropriate event / context pairs and fed to the ML model. Subsequently, during a run-time anomaly detection phase, the ML model may be used to predict a next event based on a sequence of immediately preceding events to facilitate identification of anomalous activity.

The invention discloses a landfill leakage detection method and system based on a boundary double-voltage electric field, and the method comprises the steps: building an artificial electric field which covers the whole area of a landfill along a boundary measuring line electrode at the periphery of the landfill based on the high-resistance characteristic of an impermeable film and the conductive characteristic of leachate, and collecting the boundary potential and electric field data through the high and low double-voltage artificial electric field, the method comprises the following steps: constructing a boundary survey line data pair of physical alignment through differential denoising and normalization preprocessing, adopting an encoder-decoder architecture based on 1D-CNN, fusing spatial transformation and a Focal Loss loss function, directly learning a mapping relation between a boundary survey line potential anomaly mode and an internal leakage position, and outputting a leakage probability distribution diagram of a corresponding landfill contour. The method does not need to excavate and pre-embed the landfill pile body, the positioning precision reaches + / -2m, the single detection time consumption is less than or equal to 4h, the comprehensive cost is reduced by 70%, the electrode reuse rate is greater than or equal to 90%, the method is suitable for landfill sites with various shapes and various stages, and efficient, accurate and low-cost detection of the leakage area is realized.

The embodiment of the invention discloses an automatic rock extraction method and system based on deep learning and Mars rover camera images, and aims to solve the problems that an existing Mars rock segmentation algorithm is insufficient in generalization ability under a complex earth surface background, high in model calculation load and difficult to deploy on satellite-borne edge equipment. The core of the method is that a lightweight encoder-decoder network is constructed, and the network integrates three key modules: a frequency-assisted enhancement Mama module, which accurately captures rock texture and contour by fusing the global sequence modeling capability of Mama and the frequency domain enhancement of wavelet transform; the multi-scale feature intensifier is used for adaptively fusing multi-level features by using a parallel double attention mechanism; and the boundary perception auxiliary branch improves the integrity of the segmented contour through an explicit edge supervision and feature decoupling mechanism. According to the method, the rock extraction precision is remarkably improved, meanwhile, the model complexity and the calculation overhead are greatly reduced, the method is suitable for outer space exploration scenes with limited communication bandwidth and calculation resources such as Mars rovers, and effective balance of high precision and light weight is achieved.

The invention discloses a generative adversarial network (GAN)-based recoverable watermarking resisting method, which not only can ensure that an image is protected by privacy and copyright, but also can only allow an authorized party to safely use. Specifically, an encoder-decoder-restorer architecture based on a GAN (Generic Area Network) is innovatively designed, and watermarking-resistant embedding and extraction and image restoration are established as a unified task. And a dual optimization strategy is designed to resist loss and a dynamic joint training strategy, so as to achieve ideal balance among image quality, copyright protection, privacy protection and image recovery capability. Experimental results show that the method provided by the invention can ensure that the authorized DNN classifier recovers the protected image to perform accurate classification and identification when needed, and meanwhile, excellent privacy and copyright protection capability is also provided.