223 results about "Hamming distance" patented technology

The invention belongs to the technical field of cross-modal information retrieval, and provides a quick cross-modal retrieval method and system fused with fine-grained semanteme, multi-modal data to be retrieved are obtained, a trained cross-modal retrieval model is used for learning to obtain continuous Hash features, discrete Hash codes are generated according to the continuous Hash features, and the multi-modal data to be retrieved are retrieved according to the discrete Hash codes. Performing cross-modal matching and retrieval based on Hamming distance matching, and feeding back a retrieval result; in the training process of the cross-modal retrieval model, the fusion loss function, the cross-modal alignment loss function and the shared subspace loss function are subjected to weighted fusion, and corresponding parameters are optimized with the purpose of minimizing the total loss function after weighted fusion until training requirements are met. According to the method, the cross-modal semantic consistency among different modal features is optimized, so that the generated Hash code can reflect semantic association among multi-modal data more accurately, and the precision and efficiency of cross-modal retrieval are improved.

The invention relates to the technical field of intelligent traffic, and discloses a camera state judgment method based on road surface covering and Hash comparison, and the method sequentially comprises the steps: extracting key frame pairs from a video stream at intervals; performing road surface region extraction on each frame of image to obtain a road surface mask; carrying out covering processing on the road surface area in each frame of image based on the mask to obtain a covered image; calculating a difference hash feature of each frame of covered image to obtain a hash code; calculating the Hamming distance between the Hash codes of the key frame pair; and judging whether the camera is in a rotating or stable state according to a comparison result of the Hamming distance and a threshold value. The method only depends on the video image data, does not need an external sensor, effectively eliminates the dynamic interference of the vehicle through covering the road surface, achieves the efficient and robust judgment of the state of the camera through combining the lightweight difference hash calculation, is low in calculation cost, and is suitable for edge equipment and complex road environments.

The invention relates to the technical field of data synchronization, in particular to a digestive system department clinical information acquisition and synchronization system, which comprises a multi-source heterogeneous acquisition module for generating a standardized diagnosis and treatment sequence, a characteristic spectrum construction module for mapping a text and an image into vector nodes and constructing a dynamic diagnosis and treatment characteristic spectrum, the state fingerprint verification module calculates a map fingerprint Hamming distance to position a difference feature node, and the incremental collaborative synchronization module constructs an incremental data packet and sends the incremental data packet to the central server. According to the method, the multi-dimensional characteristic spectrum based on the diagnosis and treatment time sequence is constructed, discrete images and texts are converted into topological association units, heterogeneous data semantic level alignment and integrity verification are achieved, version conflicts and information faults are eliminated, meanwhile, a dynamic hash fingerprint difference comparison strategy is adopted, accurate recognition is achieved, and only substantial change nodes are transmitted; and the network load is reduced, and high real-time consistency and zero-loss circulation of whole-flow information are ensured.

The invention discloses an unsupervised cross-modal Hash retrieval method based on multistage interaction, which belongs to the technical field of cross-modal Hash retrieval, and mainly comprises the following steps of: preprocessing cross-modal data; an unsupervised cross-modal Hash retrieval network based on multistage interaction is constructed; network training: inputting the image-text training samples into the constructed network in batches for network training; and modal retrieval: inputting the image-text query sample set and the retrieval sample set into the trained multi-level interaction-based unsupervised cross-modal Hash retrieval network, respectively generating corresponding Hash codes, and obtaining a query result by calculating the Hamming distance between the Hash codes of the query sample and the retrieval sample. The minimum Hamming distance is the final query result. According to the method, high-dimensional multi-modal features can be compressed into compact binary codes, the cross-modal retrieval efficiency is remarkably improved, and the method can be used for real-time image search and cross-modal recommendation systems.

The invention provides a multi-modal Hash retrieval method and device for an agent with a body, and relates to the technical field of artificial intelligence application, and the method comprises the following steps: constructing a multi-modal Hash retrieval model for the agent with the body; the multi-modal Hash retrieval model comprises a multi-modal feature encoder and a Kolmogorov-Arnod Hash layer which are connected in sequence; constructing a joint loss function fusing multi-modal supervision optimal transmission alignment loss and self-paced learning loss, and training the multi-modal Hash retrieval model by using the joint loss function to obtain a trained multi-modal Hash retrieval model; receiving multi-modal feature retrieval data, inputting the multi-modal feature retrieval data into the trained multi-modal hash retrieval model, and outputting a hash code of the multi-modal feature retrieval data; and calculating the Hamming distance of the Hash code of the multi-modal feature retrieval data, and returning a multi-modal feature retrieval result according to a sorting result of the Hamming distance. According to the method, the retrieval effect and the retrieval efficiency can be effectively improved.

The invention provides an SAR target screening method based on binary super-dimensional calculation, and relates to the technical field of real-time target detection. The method comprises the following steps: directly carrying out real and virtual part solution and sign function quantization on I / Q dual-channel complex data of an original echo of a synthetic aperture radar (SAR) to obtain a binary fusion matrix; constructing an abnormal profile based on a preset background statistical model, and adaptively dividing the quantized data into multiple segments according to the abnormal profile; performing super-dimensional mapping on each segment through an independent random binary projection matrix, generating segmented super-vectors, and splicing the segmented super-vectors into a complete sample super-vector; and finally, target discrimination is realized by calculating the Hamming distance between the target and a pre-stored background / target prototype super vector. The quantization logic can be dynamically switched according to the target size, and the problem that a single quantization strategy is poor in adaptability is effectively solved. And through an adaptive segmentation mechanism of energy perception, the weak target characterization capability is effectively enhanced, and the screening accuracy under a strong clutter background is improved.

A deep learning model Hash identification method based on multi-feature fusion comprises the following steps: firstly, extracting structural features (layer type sorting de-duplication), parameter statistical features (quantile normalization and histogram statistics) and functional behavior features (middle layer output quantiles) of a deep learning model; then, model structure hash, parameter hash and function hash are calculated respectively; thirdly, generating a comprehensive hash value by fusing triple hash; meanwhile, a TLSH (Trend Micro Location-Sensitive Hashing) algorithm is adopted, so that the tolerance of the TLSH algorithm to the tiny fluctuation of the parameters is improved; and finally, based on Hamming distance, calculating model similarity, and realizing efficient black box model traceability and infringement detection. The method does not need to modify a model structure or parameters, does not depend on external data, supports high robustness to attacks such as pruning and fine tuning, and is suitable for various network architectures such as CNN and Transform. The method has the advantages of lightweight calculation, high concealment and cross-platform adaptation capability, and can be widely applied to intellectual property protection and infringement tracking scenes of a deep learning model.

The invention provides an image texture feature extraction method and system based on a Hamming-Pedeson graph, and the method comprises the steps: directionally analyzing the relation between pixels of an image according to the Pedeson graph, describing the local fluctuation trend of the relation through employing a threshold function and a structural mode, constructing a local three-valued pattern PLTP feature vector of the Pedeson graph, and carrying out the feature extraction of the local three-valued pattern PLTP feature vector of the Pedeson graph. The robustness, the discrimination capability and the applicability of texture feature extraction operators are improved; the Hamming distance between the central point of the PLTP code and the adjacent point is calculated to extract the spatial structure of the PLTP code, so that the HDPLBP feature vector based on the Hamming distance is generated, and the accuracy of image texture description is improved; the statistical histograms of the PLTP descriptor and the HDPLBP descriptor are combined into the PLTPHHDPLTP texture feature histogram, so that the texture feature quantity can more intuitively and effectively represent the image texture condition, and the texture description capability is ensured to be more robust and stable. Compared with a traditional texture descriptor, image texture analysis is more effective, local information and spatial distribution information are considered, and the function of completely extracting image texture features is achieved.

The invention discloses a pilot protection method, system and device for a new energy station sending-out line, and relates to the technical field of new energy station sending-out line protection.The method comprises the steps that transient fault current signals of the new energy station side and the system side on the two sides of a fault are decomposed into different frequency components through DCT; obtaining a two-dimensional DCT coefficient corresponding to the current data at each time point, and generating a binary image; the binary images are flattened into one-dimensional vectors to obtain respective perceptual hash values, the Hamming distance between the two perceptual hash values is calculated and compared to judge the similarity of waveforms, and pilot protection is carried out according to the similarity; according to the method, fault positioning of a sending-out line and accurate action of a protection device are achieved, the method is suitable for a network-forming type large-scale new energy sending-out system, and rapid and efficient fault processing is achieved.

The invention discloses an agricultural disease hash retrieval method based on DV stabilization and adaptive feature enhancement, and belongs to the technical field of agricultural disease image retrieval and deep hash, and the method comprises the following steps: 1, obtaining an original disease image as the input of a backbone network; step 2, extracting disease image features by using a backbone network; step 3, based on a self-adaptive feature enhancement module, enhancing the disease image features to obtain self-adaptive image features; step 4, optimizing network parameters through a total loss function containing DV stabilization; step 5, obtaining a binary hash code through the hash layer; and step 6, carrying out Hamming distance sorting on the obtained Hamming codes and the Hamming codes of all the images in the retrieval set obtained by the same method, and returning a plurality of disease images with the Hamming distance smaller than a preset Hamming distance. The agricultural disease image retrieval efficiency and precision are effectively improved.

The invention relates to the field of deformation monitoring, in particular to an image deformation monitoring method, device and equipment and a program product. The method comprises the steps that matching errors are remarkably reduced by combining a Zhang's calibration method with epipolar correction, and the signal-to-noise ratio and the contrast ratio of an image are enhanced through median + wavelet filtering and histogram equalization processing. Through multi-scale census transformation and weighted Hamming distance calculation, local and global information can be effectively combined, and the robustness of a cost matrix is enhanced; according to the eight-path aggregation optimization method based on dynamic planning, the precision and the calculation efficiency of parallax calculation are further improved. According to the full-image displacement calculation method based on SIFT feature guidance, the problem of displacement propagation error accumulation is solved through sub-region matching and Gaussian Newton iterative optimization, the displacement measurement precision of a low-texture region is remarkably improved, and good robustness and accuracy are achieved in a complex scene.

The invention discloses a defect detection method for a reticle mask, which relates to the technical field of mask defect detection and comprises the following steps of: automatically calling and solidifying detection parameters according to an instruction multiplying power and a mask name; performing quadratic polynomial fitting normalization correction on the reference image according to the gray level of the light calibration point; dividing a detection area, extracting ORB features, and obtaining sub-pixel homonymy point pairs through Hamming distance and RANSAC; and performing differential defect detection after solving an affine matrix to precisely align the Die image. According to the method, manual re-judgment can be replaced, the detection time is shortened to the second level, the defect recognition accuracy is larger than or equal to 99%, and the efficient and accurate requirements of 28 nm nodes are met.

The invention relates to the technical field of big data processing and network security, and discloses a ubiquitous network data deduplication system supporting similarity data ownership verification, which comprises an encryption module, a label generation module, a deduplication judgment module, a challenge generation module, a certification calculation module, a verification module and an integrity evidence storage module. According to the system, a plaintext data block is encrypted by deriving a high-entropy session key through two-dimensional chaotic mapping, a similarity label with a threshold value is generated by adopting # imgabs0 #, and efficient de-duplication is realized based on three-section type Hamming distance judgment; a dynamic challenge-response mechanism is utilized to provide auditable ownership proof for each downloading request, and a # imgabs1 # tree and a block chain evidence storage technology are combined to realize data integrity multi-granularity verification and rapid damaged block positioning, so that storage efficiency, security and availability are considered.

The invention belongs to an infringement recognition method, and provides a three-dimensional model block chain copyright infringement recognition method, system and program product for solving the problems that an existing three-dimensional digital copyright infringement recognition method is low in recognition accuracy, poor in recognition effect in a complex scene and not applicable in a high-precision requirement scene. And calculating the Hamming distance between the copyright fingerprint set of the three-dimensional model to be detected and the copyright fingerprint set of the three-dimensional model registered on the given chain, and judging whether to infringer according to the relationship between the Hamming distance and a preset overall matching rate threshold. Wherein during copyright fingerprint set calculation, key point scales in all vertexes of a three-dimensional model are determined according to the geometric volume and surface complexity of the three-dimensional model, mixed feature descriptors are obtained by combining geometric structures and surface texture changes of local neighborhoods of candidate key points, and corresponding fingerprint abstract values are calculated after Hash processing. And the recognition stability under non-infringement operations such as rotation, zooming, slight editing and the like is remarkably improved.

The invention discloses a security and protection monitoring big data processing system based on cloud computing, and the system comprises a data receiving module which is used for receiving security and protection monitoring big data which comprises video data; relates to the technical field of data processing, through the arrangement of a target picture frame extraction module, video data are not directly transmitted, stored and processed after the video data are received, a target picture frame is extracted from mass video data, a specific area is determined by using a corner detection algorithm, and a target picture frame is obtained. A to-be-deleted picture group is divided through comparison values generated by comparing a specific area through a Bhattacharyya distance, a Hamming distance and the like, a large number of repeated or similar picture frames in the video data are deleted, then the target picture frame group is transmitted, stored and processed subsequently, the data volume of subsequent processing is greatly reduced for the video data, and the video data processing efficiency is improved. Calculation time and storage resources are effectively saved, and the overall processing efficiency of the video data is improved.

The invention discloses a multi-attribute decision-making method and system based on normal swing hesitant fuzzy information and a foreground theory. The method comprises the following steps: constructing an initial decision-making matrix and carrying out normalization processing; generating a multi-dimensional hesitant fuzzy decision matrix by calculating the Hamming distance, the chi-square distance and the trigonometric function distance between each scheme and the positive ideal solution, and converting the multi-dimensional hesitant fuzzy decision matrix into a normal swing hesitant fuzzy decision matrix; the comprehensive hesitant fuzzy entropy method and the logarithm percentage change driven target weighting method are fused to determine the comprehensive weight of the index; mapping the negative ideal solution distance into an income value and mapping the positive ideal solution distance into a loss value in combination with a foreground theory, and generating a foreground value decision matrix and normalizing the foreground value decision matrix; and calculating the correlation degree between the scheme and the dynamic ideal solution through a normal swing hesitant fuzzy weighting correlation coefficient, thereby realizing scheme sorting. According to the method, various kinds of uncertainty information and incomplete psychological characteristics of decision makers can be fully considered, and more accurate index weights and decision results are provided.

An authentication method includes: computing, in a secret state, a hash value from a first feature value of the biometric information, using a hash function; converting the hash value from a first integer share to a binary share; computing, in a secret state, an XOR bit sequence by performing an exclusive OR operation between the hash value and a registered hash value; converting the XOR bit sequence from the binary share to a second integer share; computing, in a secret state, a total of a plurality of bit values present in the XOR bit sequence; and determining, using the Hamming distance, whether a registered feature value is to be used in authentication processing.

The invention discloses a smart park comprehensive management system based on intelligent collection, and relates to the technical field of smart park comprehensive management, and the system comprises a collection and generation module which collects original data, segments a service area through a Voronoi graph, and generates a shortest distance matrix in combination with a Floyd algorithm; the calculation and iteration module is used for calculating facility priority scores, generating a facility start-stop sequence and scheduling individuals in combination with the shortest distance matrix, and carrying out iterative optimization to obtain an initial solution; the optimization module is used for generating an elite and a common population based on the initial solution to perform initialization, and performing optimization through disturbance, path selection probability and a pheromone diffusion mechanism in combination with a Hamming distance difference evaluation and cross replacement strategy to obtain an optimal solution; according to the invention, the position of the facility can be accurately adjusted, the facility arrangement of the park is more reasonable, and the service range is maximized, so that the overall service quality and the operation efficiency are improved.

The invention discloses a network fast security verification method and system. The method comprises the steps that an initial device generates a verification request and records a corresponding request timestamp; collecting a first device feature to perform migration feature calculation to obtain a migration feature classification result, generating a first feature hash code based on the migration feature classification result, and generating an initial feature hash code set in combination with the request timestamp; determining a migration frequency and a migration state according to the initial feature hash code set, selecting a corresponding verification process based on the migration frequency and the migration state, and when the migration frequency and the migration state are not changed, performing first verification based on the Hamming distance between the first feature hash code and the initial feature hash code set; and when the migration times and the migration state change, generating a second feature hash code by using the migration token, the second device feature and the verification times, and performing Hamming distance comparison on the second feature hash code and a third feature hash code generated according to the initial feature hash code set and the verification times to realize efficient device identity verification.

Methods for analyzing a nucleic acid sequence are described, including methods for generating a sequencing coding scheme that accommodates for errors during the sequencing process. The methods may comprise, for example, generating different codewords; assigning a different codeword to each one of four different bases, wherein the different codewords adhere to designated Hamming distance rules; contacting different polynucleotide templates with nucleotides of the four different bases; allowing binding and optional incorporation of the nucleotides; imaging the different polynucleotide templates to determine recorded signal code sequences; and comparing recorded signal code sequences to the different codewords to identify bases corresponding to the recorded signal code sequences and / or checking for errors in the recorded signal code sequences.

The application belongs to the technical field of new energy sending-out line protection, and provides a new energy station sending-out line pilot protection method and system based on Hamming distance. The new energy station sending-out line pilot protection method comprises obtaining current sampling values of the station side and the system side of the sending-out line after protection starting; data preprocessing is performed on the current sampling values of the station side and the system side of the sending-out line, and a first numerical sequence and a second numerical sequence containing 0 and 1 are obtained respectively; the Hamming distance is calculated according to the ratio of the different sampling point numbers between the first numerical sequence and the second numerical sequence to the sampling point number of the sampling window length; the fault type is judged according to the size of the Hamming distance, and whether to issue a trip command to the protection devices of the station side and the system side of the sending-out line is judged in combination with the comparison result of the Hamming distance and the setting value of the Hamming distance.

To improve the tamper resistance of semiconductor devices against physical attacks. [Solution] The semiconductor device according to this disclosure includes a transition circuit that transitions a state at a predetermined Hamming distance according to the result of comparing a value to be authenticated with a reference value, and a determination circuit that determines the validity of the authentication by determining whether the Hamming distance between the state before the transition and the state after the transition matches the predetermined Hamming distance.

The present invention relates to the field of data storage, and discloses a DNA-based dynamic balancing system, a data storage method, and a decoding method. During the process of encoding the first data to obtain a DNA molecular chain, various constraints are added to the first address and the second address, enabling the encoded data to be read efficiently and accurately. For example, the Hamming distance between the first address and the second address is greater than or equal to half of the length of the first address, reducing the possibility of address selection errors during reading; the prefix of the first address is different from the prefix of the second address and the suffix of the second address, avoiding the possibility of matching errors during the reading process; the content of guanine and cytosine in the prefix of each primer accounts for a preset ratio of the total content of guanine, cytosine, adenine, and thymine contained in the primer, resulting in high accuracy when sequencing is required to read the encoded data in advance.

The application discloses a keyword extraction method, system, device and storage medium. The method comprises the following steps: obtaining a sensitive word group list, removing stop words and splicing words in each sensitive word group, and obtaining preprocessed text; based on the preprocessed text, a word vector is constructed, and a mapping relationship between a word group in the preprocessed text and the splicing words is established; an initial cluster center vector is selected in the word vector, a Hamming distance between a remaining word vector and the initial cluster center vector is calculated, the word vector is divided into clusters according to the Hamming distance, and a preliminarily divided keyword cluster is obtained; based on the word vector and the preliminarily divided keyword cluster, an adaptive clustering algorithm is adopted to obtain a similar word vector cluster; and the similar word vector cluster is spliced with the splicing words according to the mapping relationship, so that a keyword list is generated; and it is ensured that the obtained keywords can accurately reflect the characteristics of the text content.

The invention discloses a super-dimensional vector similarity query-oriented high-energy-efficiency coarse-grained in-memory query method, which comprises the following steps of: segmenting a super-dimensional vector, performing query operation on each segment of input super-dimensional vector, and calculating a Hamming distance between the input super-dimensional vector and the corresponding segment of stored super-dimensional vector; comparing each section of input super-dimensional vector with a storage section step by step, and compressing the Hamming distance between each section of super-dimensional vector into a binary code with a fixed length; and summarizing the compressed codes of the Hamming distance of each segment, and calculating the matching degree between the super-long-dimension input feature vector data and the storage category data by fusing the compressed Hamming distance information of the input super-dimension vector of each segment and the stored super-dimension vector of the corresponding segment. According to the method, the matching degree between the super-long-dimension input feature vector data and the storage category data can be efficiently calculated, so that low-complexity coarse-grained super-dimension vector similarity query is realized.

The present application relates to the technical field of cross-modal retrieval, and particularly relates to a strong correlation unsupervised cross-modal retrieval method guided by information quantity, which is realized through the following steps: firstly, image local features and global features, and text features are extracted; the image local features and global features are enhanced; the enhanced local features are subjected to regularization processing; then, the image global features and local features are subjected to orthogonal fusion using an image feature fusion network; next, the image features and text features are fused according to a different modal feature information quantity conversion ratio principle using a multi-modal fusion network; finally, different modal features are mapped into hash codes, and Hamming distance is used for similarity sorting, so as to obtain a retrieval result. The present application focuses on the enhancement and fusion of data features, can obtain more semantic information, and improves the retrieval efficiency.

The invention discloses a novel three-dimensional laser point cloud feature extraction and matching method. The method comprises the following steps: obtaining a three-dimensional point cloud through radar scanning; dividing point cloud units, and extracting point cloud feature key points of static blocks, vertical edges and ground contours; a key point is randomly selected, the key point is taken as a body center, and a three-dimensional longitudinal segmentation mode is adopted to divide the three-dimensional point cloud; three main directions are selected as initial positions, and the shortest distance and the distribution density value between a unit where the key point is located and an anticlockwise adjacent unit are calculated in sequence to serve as descriptors of the key point in the three main directions; performing binary transformation on the descriptors, and integrating the descriptors in an XOR mode to obtain descriptors; and on the basis of the Hamming distance of the CPU instruction set, feature matching between radar frames is calculated, so that instant positioning and regional map construction of the mobile robot are realized. Experimental results show that the calculation efficiency of the method is higher than that of Link3D, and the method shows more excellent estimation precision and robustness than other odometer frames.

The invention belongs to the technical field of face recognition, and particularly relates to a batch face analysis and comparison method and system. The method comprises the following steps: extracting key features in face data, namely a two-eye pupil distance, and performing batch calculation by utilizing a Hamming distance algorithm according to the two-eye pupil distance in the face data to obtain a plurality of pieces of face data with the highest similarity, and then, on the basis of a plurality of pieces of acquired face data with the highest similarity, performing similarity comparison on all features in the face data by adopting a cosine algorithm, and acquiring the same face data according to a similarity comparison result. The system comprises a face data snapshot module, a key feature extraction module, a face similarity extraction module and a face confirmation module which are connected in sequence. Compared with original batch comparison, the method and the system are faster, more efficient and more intensive.

The invention provides an automatic component classification method oriented to multi-level category labels and related equipment. Comprising the following steps: preprocessing to-be-identified component data with a plurality of fields; extracting a text feature in the to-be-identified component data with the first format, and fusing the hierarchical structure information of the category label into the text feature to obtain a fused feature vector; loading a preset coding matrix based on the DECOC algorithm; the coding matrix is provided with a plurality of corresponding binary classifiers; fusing prediction probabilities of a plurality of pre-trained base classifiers based on a soft voting mechanism in each binary classifier to obtain corresponding weighted prediction probabilities so as to obtain a plurality of predicted weighted probabilities corresponding to the plurality of binary classifiers; combining the plurality of weighted prediction probabilities into a prediction code word; and determining a prediction category label group according to the prediction code word and the Hamming distance of each row in a preset coding matrix. And the types of the components can be quickly and accurately predicted.

The invention provides a ciphertext classification model anti-attack method introducing error correction codes. The method comprises the following steps: performing redundant information addition processing on original ciphertext data through an ECC (Error Correction Code) scheme; inputting the current adversarial sample into the target ciphertext classification model for forward propagation, and then calculating a model output result and a loss value of a real encryption algorithm label; solving the embedding gradient of the loss value relative to the current adversarial sample through back propagation; combining gradient income, cost and ECC residual error correction quota to construct a comprehensive evaluation index, and screening an optimal character for replacement to generate a candidate confrontation sample; the disturbance amplitude of the candidate adversarial sample is verified through the Hamming distance, and if the disturbance amplitude exceeds the maximum ECC error correction capability, discrete ECC safety projection is executed for constraint, and a new adversarial sample is obtained; and the confrontation sample with successful deception model is verified at the receiving end. According to the method, the integrity and the concealment of communication data are kept while the attack is effectively resisted.