312 results about "Information recovery" patented technology

The invention provides an abdomen multi-organ image segmentation method fusing multi-scale features, and the method comprises the steps: S1, constructing a space detail fusion module extraction network, which comprises a space attention module, a channel attention module, an up-sampling operation, a down-sampling operation and a feature smoothing operation; s2, designing an encoder, and connecting the encoder with the space detail fusion module; s3, transmitting a plurality of features with different resolutions obtained by the encoder to a decoder through jump connection for information recovery; and S4, arranging the abdominal multi-organ medical image data set to perform model training, and performing multi-organ segmentation prediction on the abdominal multi-organ medical image. According to the method, local and global information is fully utilized, multi-scale feature information is combined, the image feature extraction and expression capability of the model is improved, and the segmentation precision of multiple organs of the abdomen is improved.

The invention relates to the technical field of crop yield analysis, in particular to a crop yield prediction method based on a time sequence remote sensing image, and the method comprises the steps: obtaining an optical remote sensing image and a synthetic aperture radar image in a crop growth period, and constructing a time sequence image through complementary information fusion and under-cloud information recovery; the problems of cloud interference and data missing are solved, a multi-scale time regulation and control convolution module is used for processing a time sequence image, multi-scale time sequence features are extracted, feature fusion is carried out in combination with attention difference jump connection, and a rice distribution map is obtained; determining a rice planting area based on the rice distribution diagram, selecting an optimal assimilation time period in the rice planting area, and assimilating the leaf area index subjected to remote sensing inversion and a crop growth model simulation result to generate a simulated yield after assimilation; based on the multi-scale time sequence characteristics, a parameter dynamic response matrix related to crop growth model parameters is established, and residual error correction is performed on assimilated simulation yield, so that the rice yield prediction precision is improved.

The application discloses a ciphertext decryption method and related equipment applied to ciphertext encrypted according to an RSA algorithm, and the method comprises the following steps: acquiring a large number in ciphertext and a public key of the ciphertext; selecting a number smaller than the large number as a base of a discrete logarithm problem, selecting 1 as a modulus power operation result corresponding to the base, and converting the base and the modulus power operation result into an initial quantum state; inputting the initial quantum state into a quantum computing module for solving the discrete logarithm problem to obtain a first target quantum state in which a solution result of the discrete logarithm problem is stored; extracting the solution result from the first target quantum state, and calculating a private key of the ciphertext based on the solution result; and decrypting the ciphertext based on the private key to obtain plaintext of the ciphertext. Through the technical scheme, the private key can be recovered according to the disclosed information in the case that the private key is not disclosed, and then the ciphertext is decrypted, so that the related technical blank is filled.

The invention provides a semantic communication method and system for a high-resolution image, and relates to the technical field of semantic communication, and the method comprises the steps: obtaining a semantic feature and super-prior of a to-be-coded image; segmenting the semantic features into four groups of feature elements; based on the four groups of feature elements, the hyper-priori and the channel state, four-step airspace priori and semantic importance are obtained; based on the feature element combination corresponding to the ith estimation index, the ith semantic importance, the semantic feature coding result of each path before the ith path and the channel state, utilizing a semantic feature codec to obtain received semantic features; and based on the received semantic features and channel states, utilizing a semantic information restorer and a denoising enhancement module to reconstruct an image. The semantic features of the image are extracted and recovered by adopting the constructed module, and the semantic features of the image are processed by utilizing a coder-decoder, so that the coding and decoding efficiency is improved, the understanding ability of the model to the image content is enhanced, and the high-quality image content can be better reconstructed.

The invention provides a hot upgrade method for a virtual machine monitor (VMM), which comprises the following steps of: pausing the operation of a virtual machine, storing state information of an original VMM, and reserving a kernel-based virtual machine KVM extension page table and a memory management structure of the original VMM; reserving a page global directory and a virtual memory area VMA corresponding to a virtual machine RAM area according to the memory management structure, and refreshing a process address space described by the memory management structure into a new VMM execution program so as to complete switching of the original VMM into the new VMM; and executing the new VMM, recovering the state information to the new VMM, and inheriting the KVM extension page table by the new VMM to recover the operation of the virtual machine. The invention further provides computer equipment, a computer readable medium and a computer program product.

The invention provides a large model-based text generation and information recovery method and system. The method comprises the following steps of: converting complex situation information into simple and accurate text description by utilizing a text generation model; converting the text description of the situation information into a semantic vector by using a text encoder of a pre-trained CLIP model, and capturing semantic information of the text; encoding the noise image into a hidden space representation by using an encoder of a variational auto-encoder, and learning a compressed representation of the image; fusing the semantic vector generated by the pre-trained CLIP model with the image hidden space representation generated by the encoder of the variational auto-encoder, and optimizing the fused hidden space representation by using the trained de-noising Unet network; the optimized hidden space representation is decoded into a high quality situational information image using a decoder of a variational auto-encoder. By applying the technical scheme of the invention, the technical problem of incomplete received information caused by inevitable information loss in the data compression process in the prior art is solved.

The invention discloses a screen shooting traceability method, device and equipment and a readable storage medium, which are applied to the field of computer software, and comprise the following steps: encoding watermark information to obtain a watermark information sequence; embedding the watermark information sequence into the image through an R channel, a G channel and a B channel of RGB (Red, Green, Blue) to obtain an image with a watermark; and acquiring the shot image with the watermark image, and tracing to obtain watermark information based on the shot image. According to the method, the watermark is generated by using the RGB channel of the single pixel image, so that the visual concealment of the watermark image is improved, the difficulty in perceiving of human eye recognition is ensured, and the visual quality of a user is not influenced; the watermark embedding position is not influenced by an original image carrier, and watermark information can be hidden in any area of the image; and the watermark information recovery success rate is high.

The invention relates to a multi-temporal optical remote sensing image thick cloud removal model, method and device. The model comprises a feature extraction module, a feature fusion module and an image reconstruction module which are connected in sequence. The feature extraction module is used for sequentially performing nonlinear feature extraction, channel number adjustment and multi-scale feature extraction on the multi-temporal satellite image to obtain multi-scale features; the feature fusion module is used for sequentially performing scale compression, summation operation and self-attention-based convolution modulation operation on the multi-scale features to obtain global information; and the image reconstruction module is used for performing up-sampling and affine transformation on the global information in sequence to obtain modulation features, and performing low-frequency information-to-high-frequency information processing and high-frequency information reconstruction processing on the modulation features in sequence to obtain a cloudless image. According to the method, the spectral information recovery and semantic information retention capability under thick cloud shielding is improved, and the cloud removal performance is further improved.

The invention provides an RAID (redundant array of independent disks) data processing method and device, a chip, electronic equipment, a storage medium and a computer program product, and is applied to RAID, the method comprises the following steps: detecting that at most three storage devices fail, determining at most three corresponding to-be-recovered areas in a stripe, and storing the to-be-recovered areas in the stripe; the to-be-recovered area is a data area, a verification area, a first hot standby area and / or a second hot standby area; determining a target area based on at most three areas to be recovered, wherein the target area is a data area, a verification area, a first hot standby area and / or a second hot standby area except the areas to be recovered; and target information in the target area is read, and the data information and / or the first verification information of the to-be-recovered area are / is recovered based on the target information.

The invention provides an emergency communication network generation method, a terminal device and an electronic device, the method is applied to a first node device in a wireless communication network, and when the communication of the wireless communication network is interrupted due to unforeseen circumstances, each node device monitors a neighbor node list broadcasted by other node devices to generate an emergency communication network. And comparing the received neighbor node list with the neighbor node list of the local machine, determining cluster head equipment, and forming a corresponding target communication cluster by taking the cluster head equipment as a central control node. And the cluster head device sends a cluster establishment message to the effective neighbor node in the neighbor node list, the effective neighbor node adds to the target communication cluster as the identity of the cluster member based on the received cluster establishment message, and finally each cluster member device carries out data receiving and sending according to the unified scheduling information sent by the cluster head device. Therefore, each node device can quickly form a communication cluster and then quickly recover communication according to the unified scheduling information sent by the cluster head device.

A mesh data decoding device according to embodiments may comprise: a reception unit for receiving a base mesh bitstream, a displacement vector bitstream, a texture map bitstream, and signaling information; a base mesh processing unit for reconstructing a base mesh from the base mesh bitstream; a displacement information processing unit for reconstructing displacement information from the displacement vector bitstream; a reconstruction unit for reconstructing a mesh on the basis of the base mesh and the displacement information; and a texture map processing unit for reconstructing a texture map from the texture map bitstream.

The invention relates to a data recovery method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: receiving a data reading command, and determining protection information of target data corresponding to the data reading command; the APP domain information in the protection information comprises bad sector mark BST information corresponding to a sector where the target data is located; under the condition that the BST information is damaged, verification information corresponding to the APP domain information is determined in the APP domain stripe; verifying the APP domain information according to the verification information to obtain recovered APP domain information; the recovered APP domain information contains the recovered BST information. By adopting the method, the data recovery efficiency can be improved.

The embodiment of the invention provides a low-illumination image enhancement method and system based on frequency decomposition and multi-scale fusion, and the method comprises the steps: obtaining a to-be-processed image, and decomposing the to-be-processed image into a low-frequency component feature map and a high-frequency component feature map through employing an iterative mean filter; a low-frequency tensor is generated based on the low-frequency component feature map, and a high-frequency tensor is generated based on the high-frequency component feature map. And fusing the high-frequency tensor and the low-frequency tensor to generate an enhanced feature map. According to the method, iterative mean filtering can be applied to separate frequency bands, and an image to be processed is decomposed into a high-frequency component and a low-frequency component. More accurate image enhancement is realized through different sub-networks, that is, a high-frequency enhancement network focuses on detail reconstruction and noise suppression, and a low-frequency enhancement network is responsible for recovery and transformation of illumination and color information, so that the image enhancement quality can be improved, and the problem of poor image enhancement effect is solved.

The invention discloses a CSI (Channel State Information) feedback method and system based on deep learning, belongs to the technical field of communication, and aims to solve the technical problems of low recovery precision, reduced frequency efficiency and huge uplink bandwidth consumption of CSI feedback in a large-scale MIMO (Multiple Input Multiple Output) system. Comprising the following steps that a channel state information feedback model TCMRNet is constructed and trained, the TCMRNet comprises an encoder and a decoder, LSTM network units and convolution layers with different convolution kernel sizes are introduced into the encoder and the decoder, the output end of each convolution layer is connected with a batch normalization layer, and the output end of each convolution layer is connected with an LSTM network unit; the encoder is used for carrying out multi-resolution feature extraction and time correlation analysis on input channel state information and compressing extracted features into feature vectors to be output, and the decoder takes the feature vectors as input and is used for decompressing the feature vectors, recovering the channel state information and outputting reconstructed channel state information; and performing time domain transformation on the reconstructed channel state information.

The invention discloses a geomagnetic anomaly data stable downward continuation method based on an SRGAN deep learning network, and belongs to the field of geophysical magnetic field data continuation. According to the invention, in combination with the continuation characteristic of geomagnetic anomaly data, a CBAM space attention mechanism module is added to improve the spatial perception capability of a network model to the magnetic anomaly data, and in combination with the characteristic that the downward continuation of potential field data can amplify the noise generated by a high-frequency signal; and the deep learning model is combined with adaptive threshold multiple wavelet transform to suppress noise data so as to further improve the continuation precision. In order to enable a model to fully learn detailed information of geomagnetic anomaly data, a training data set construction method combining a Kriging super-resolution algorithm with a sliding subarea mode is provided. By designing and synthesizing a simulation model experiment and comparing actual measurement data with other three algorithms, the model is highest in calculation precision and best in magnetic anomaly detail information recovery.

The invention relates to the technical field of over-the-air downloading, in particular to a remote OTA upgrading method and system, and the method comprises the steps: building an encrypted communication channel between a terminal and a server, and guaranteeing the precise and safe transmission of data; the terminal transmits version related information to the server through the communication channel, the server performs upgrade judgment according to the received information and stores the information as a backup for specific analysis of incremental files, and upgrade failure caused by data loss in the running process of the terminal is prevented; if upgrading is needed, the server compares the files to generate a class of file packages, and analyzes modular files in the incremental files to reduce the size of the file packages; after the file package is transmitted to the terminal, the completion of upgrading is confirmed through integrity verification and testing; in addition, when upgrading fails, the terminal can be recovered by utilizing backup information through a rollback mechanism, so that the upgrading process is accelerated, and the upgrading success rate is improved.

A mesh data decoding method according to embodiments may comprise: a step of receiving a base mesh bitstream, a displacement vector bitstream, a texture map bitstream, and signaling information; a base mesh processing step of restoring a base mesh from the base mesh bitstream; a displacement information processing step of restoring displacement information from the displacement vector bitstream; a restoration step of restoring a mesh on the basis of the base mesh and the displacement information; and a texture map processing step of restoring texture maps from the texture map bitstream.

The invention provides a communication method and a communication device. The method comprises the following steps: a first communication device compresses N pieces of second channel information based on X pieces of first channel information to obtain M pieces of compressed information, for example, the ith compressed information is obtained by compressing Pi pieces of second channel information based on the jth piece of first channel information, and the acquisition time of the first channel information is earlier than the acquisition time of the second channel information; the first communication device sends the ith compressed information, and the first communication device also sends the jth first channel information; the second communication device restores Pi pieces of second channel information based on the i-th piece of compression information and the j-th piece of first channel information that may be received at different times. Through the mode, the channel information (namely the second channel information) can be fed back in time, and the feedback overhead can be reduced.

The invention discloses a covert communication system based on a generative diffusion model, and belongs to the technical field of covert communication. In order to solve the problems of insufficient concealment and poor anti-noise performance in the prior art, hidden transmission and reliable recovery of information are realized by constructing a diffusion generation and classification discrimination module. The sending end carries out noise addition and shaping on the original bit waveform by using a forward diffusion process to form a noise-like hidden signal; and a receiving end recovers a signal structure through a reverse denoising process, and accurately judges bit information through a classifier. According to the method, a Gaussian approximation hidden waveform library is constructed by adopting a signal pair training model which is consistent in statistics but distinguishable in structure, and random sampling mapping and boundary smooth splicing are combined, so that transmission signals are difficult to distinguish from noise in a frequency domain and a feature space; therefore, the avoidance capability, the anti-interference robustness and the information recovery accuracy under the low signal-to-noise ratio of the detection means are remarkably improved.

The invention provides a gradient matching attack method of a Vision Transform (ViT) model aiming at noise disturbance, and relates to a privacy protection problem in a federated learning scene. The method specifically comprises the steps of providing a gradient denoising method based on a diffusion model, recovering disturbed gradient data through reverse noise prediction, and designing a gradient diffusion denoising network for attention perception to solve the problem of feature distortion caused by traditional Gaussian noise. On the basis, a gradient matching attack method for ViT is provided, and high-fidelity semantic information recovery of noise-added gradient is realized by analyzing gradient distribution characteristics of a ViT model and combining a position coding gradient extraction and optimization reconstruction technology. The method disclosed by the invention reveals the vulnerability of the ViT model with noise disturbance when facing gradient leakage attacks, provides a confrontation view angle and thinking enlightenment for constructing a ViT privacy protection mechanism, and has important application value in privacy sensitive fields such as medical image analysis and financial risk control.

The invention relates to the field of laser communication, in particular to a near sea surface-oriented low-altitude long-distance laser communication system and a near sea surface-oriented low-altitude long-distance laser communication scheme. A transmitting end uses an optical module to perform multi-path multi-wavelength optical OOK signal transmission, and each path of optical signal is subjected to independent transmission power control and independent transmission optical aperture transmission; a receiving end uses a large-aperture receiving optical unit for receiving, uses a large-core-diameter multimode optical fiber to replace a single-mode optical fiber to realize signal receiving with higher optical power, and uses a multimode wavelength division multiplexing demultiplexer to separate received multiple paths of optical signals. Each path of separated signals is subjected to signal detection and information recovery through a receiving optical module, and finally, the information transmission quality is judged according to the bit error rate.

The invention relates to the technical field of computer vision, and provides a spatial three-dimensional structure information recovery method and device, equipment and a storage medium, and the method comprises the steps: obtaining a target scene image shot by a color camera and internal parameters of the color camera; inputting the target scene image into a depth estimation model, and estimating scale-free depth information of the target scene image through the depth estimation model to obtain a scale-free depth map; performing scale recovery on the scale-free depth map to obtain a scale depth map; performing instance segmentation on the target scene image to obtain an instance segmentation map corresponding to the target scene image; according to the pixel coordinate position, associating the target scene image, the scale depth map and the instance segmentation map to obtain an association map; based on the internal parameters, back-projecting the association graph to the point cloud space to obtain preliminary spatial three-dimensional structure information; the three-dimensional structure information can be recovered by adopting a scheme of combining a common color camera with an algorithm, and the cost is low.

The invention discloses a camera calibration method. The method comprises the following steps: constructing an orthogonal binary stripe target with periodic characteristics; fourier transform is carried out on the image, horizontal and vertical phase shift fringes are separated from the image, the position of a feature point is positioned through a phase diagram, and pixel coordinates of the feature point are extracted; the invention further provides a method for recovering the three-dimensional shape information of the object. The method has high robustness for camera defocusing, is suitable for cameras with small depth of field, and can expand the original depth of field of the cameras.

The invention discloses a neural network structure search method for self-supervising image noise reduction, and belongs to the field of neural network structure search and image noise reduction. According to the method, a search space of a multi-scale encoder is designed, so that a network structure can fuse feature information extracted from different scales, more image details can be recovered, and the network structure with a better noise reduction effect can be found; based on a search space of a multi-scale encoder, an improved genetic algorithm is provided to search a network structure, a coarse-grained population initialization method, a distance-guided parent selection method and a modularization-based crossover mutation operation are designed, and the search capability of the genetic algorithm is improved. Compared with an existing self-supervised image noise reduction method and a neural network structure method, a better noise reduction effect can be achieved; experimental results prove that the self-supervised image noise reduction neural network structure searching method can effectively discover a network structure with a better noise reduction effect.

The invention relates to a recoverable visual encryption method for a human body privacy area in an image collected by an unmanned aerial vehicle. The method comprises the following steps: constructing a joint deep learning model integrating a pre-training target detection module, a self-adaptive visual encryption module and an enhanced visual decryption module; in the training stage, target detection module parameters are frozen, and an encryption and decryption module is optimized end to end through combined loss formed by a composite decryption loss function and an encryption strength loss function; during application, the target detection module accurately locates a human body sensitive area, the encryption module applies size-adaptive and strength-learnable pixelated transformation to generate a natural encryption image, and the decryption module realizes high-fidelity original information recovery through an enhanced U-Net network under an authorization condition. Compared with an existing traditional image encryption or irreversible desensitization technology, the method has the advantages that seamless integration with a mainstream target detection process is achieved, the integrity and restorability of information are guaranteed while high visual concealment is achieved, and the method has the advantages of being intelligent, high in efficiency and high in practicability.

The invention provides a database log analysis and recovery method and device, equipment, a medium and a program product, and relates to the field of data processing. Periodically generating a snapshot file of cache record data involved in a database log analysis process, and recording a snapshot identifier corresponding to the snapshot file; updating version information of a data dictionary, and binding the snapshot identifier with the version information of the data dictionary; when database log analysis is interrupted, matching corresponding target version information from the data dictionary according to a target snapshot identifier; recovering the state of the cache record data based on the target version information, and determining the initial position of database log analysis; according to the method, the snapshot file of the cache record data is periodically generated, and the snapshot identifier and the data dictionary version information are bound, so that the cache data state and the analysis starting position are quickly recovered based on the latest snapshot after the database log analysis is interrupted, and the fault tolerance and the recovery efficiency of log analysis are improved.

The invention provides a communication method and a communication device, and the method comprises the steps that a second device obtains first indication information from a first device, the first indication information is used for determining a time unit in which reset state information of a first artificial intelligence AI model takes effect, and the second device can determine according to the first indication information, and a time unit for enabling reset state information of the first artificial intelligence (AI) model. According to the scheme of the embodiment of the invention, the robustness of the recovery performance of the channel information can be improved.

The invention provides a semantic communication method and system for an ultra-high-definition video, and relates to the technical field of semantic communication, and the method comprises the steps: obtaining semantic features based on a to-be-coded video P frame and a channel state through employing a video P frame semantic feature extractor; based on the semantic features, obtaining four-step semantic importance by using a quadtree-based spatial domain entropy model; based on the semantic features and the four-step semantic importance, utilizing a four-way semantic feature codec to obtain the received semantic features; and reconstructing a video P frame by using a semantic information restorer and a de-noising enhancement module based on the received semantic features, the received channel state and the multi-scale reference context features. According to the method, the semantic features of the video frames are extracted and recovered by using the module constructed by the CNN, and the semantic features of the video frames are processed by using the coding and decoding module constructed by the Swinin-Transform, so that the coding and decoding efficiency is improved, the understanding ability of the model to the video content is enhanced, the high-quality video content can be better reconstructed, and the size of the model parameter is controlled.

The invention provides a model reasoning method and device, a storage medium and computer equipment. The method comprises the following steps: acquiring a pre-stored allocation table and an address recovery table of a target model; wherein the allocation table is a target model used for recording an execution sequence corresponding to nodes in a computational graph captured by the target model in a preprocessing process, and the address recovery table is a mapping relation table of a kernel and parameters constructed according to the computational graph; according to the execution sequence in the allocation table, an address recovery table is inquired, address information of nodes in the computational graph is obtained, and the address information comprises a kernel address and a parameter address of each node; according to the address information of each node, recovering a computational graph of the target model in the reasoning process; and based on the calculation graph, executing reasoning calculation of the target model. Therefore, the model can be inferred according to the computational graph, the operator graph is prevented from being recaptured when the model is inferred, the time delay of cold start of the model is reduced, and the cold start efficiency of the model is improved.

The invention discloses a single-photon single-pixel imaging method based on average pulse number estimation intensity. The method comprises the following steps: recording each single-photon event; averaging the recorded pulse number of each single photon event to obtain an average pulse number of different photon numbers; performing intensity estimation by using the average pulse number of different photon numbers to obtain average pulse number estimation intensity, and performing information recovery on the target object to obtain an image of the target object; based on the image of the target object, the imaging quality of the target object is evaluated by calculating different image quality indexes. An image is restored by counting the number of pulses before each photon arrives and averaging. In an experiment, only one single-photon avalanche diode is used as a single-photon detector to obtain photon information, the high efficiency of the method under the low-light condition is proved, and compared with a traditional time correlation single-photon counting imaging method, the method can achieve information recovery of a target object through fewer photons.