70results about How to "Guaranteed recognition efficiency" patented technology

The invention provides a design method of an active defense system based on white list technology and a system, belonging to the technical field of safety. According to the design method of the activedefense system, the reliability of a white list source is guaranteed through collecting the Hash value of a legal copy operation system mirror image; network visits are reduced through generating andcustomizing a local white list library for a customer, the recognition efficiency of a program is improved to some extent, and the starting speed of a run program is guaranteed; the pragmaticality and recognition efficiency of the white list are guaranteed through acquiring a patch updated by an operating system in real time; a recognition function is trusted through an unknown program, the resource requirement on traditional white list technology is solved, and the non-limitation of the white list is guaranteed. The method and the system can be used for effectively solving the problem whicha blacklist confronts, and are reliable, efficient, pragmatic and non-limited in the realization of the active defense system design of the white list.

The invention relates to a method and a system for identifying the bar code of a server-assisted mobile terminal, belonging to the technical field of the mobile communication value-added services. The method comprises the following steps of: collecting bar code images, resolving the bar codes, invoking the server-assisted resolving capability, transmitting through the network connection, managing the network connection, assisted-resolving the bar codes, processing resolution results, and finally showing identification results to a user at the bar code identifying client. The invention provides a two-stage identification method including the mobile terminal direct identification and the server-assisted identification. The method provides a two-stage identification method consisting of mobile terminal direct identification and server-assisted identification and comprising the following steps of: firstly, using a camera of the mobile terminal to directly scan the bar codes for identification, and judging whether to invoke the server function for assisted identification on the basis of the identification results, thereby ensuring the identification efficiency of the terminal bar codes and simultaneously achieving the goals of improving the identification success ratio and the stability. Therefore, the implement of the method is practical and feasible.

The invention relates to an image number character recognition method. An algorithm structure is optimized, and influences of illumination and inclination are processed in an image preprocessing stage, thereby ensuring that a better binary character image is acquired before character recognition; then, a two-dimensional PCA algorithm is adopted to extract primary characteristics in an image so as to reduce the dimension and operation complexity of the image; a decision tree algorithm and an Adaboost algorithm are combined to optimize a classification algorithm, thereby ensuring the recognition efficiency and reducing the over-fitting risk; and finally, a nine-one division method is used for verifying the generalization ability of a classification model so as to ensure the stability of a classifier. By virtue of algorithm optimization, the accuracy rate and speed of number character recognition are obviously increased.

The invention provides a micro-expression recognition method based on the main direction of optical flow, which comprises the following steps: interpolating a micro-expression sequence formed by framing a micro-expression video; positioning the eyes and nose of each frame, and calculating the position of the middle point of the eyes as the origin of the coordinate axis; calculating the distance dem and dmn from the origin of the coordinate axis of each frame to the center of each eye and the nose; constructing normalized mesh according to dem and dmn; calculating the optical flow direction ofeach pixel region in all adjacent frame grids and clustering to obtain the optical flow main direction; a rectangular characteristic region being formed in the main direction of the optical flow, andthe pixel region in the main direction of the optical flow being assigned with weight coefficients to obtain the optical flow field optimized by adjacent frame grids, and the curves formed by connecting the related data drawing points being smoothed to obtain the global optical flow field; the pixel region in the main direction of the optical flow being assigned with weight coefficients. A training classifier inputs each optical flow characteristic region and compares the sum of similarities between the traversed-level microexpression sequence and the trained microexpression sequence with S toobtain a classification result.

The invention discloses a voiceprint recognition method and a device, computer equipment and a storage medium, and belongs to the technical field of artificial intelligence, and the method comprises the steps: obtaining a to-be-recognized voice, carrying out the framing of the to-be-recognized voice, and obtaining an audio frame set; performing Fourier transform on the audio frame to obtain sequence information of the voice data; inputting the sequence information into a pre-trained self-attention voiceprint recognition model, and obtaining a voiceprint feature matrix of the audio frame; and splicing the voiceprint feature matrixes to obtain a voiceprint recognition result of the to-be-recognized voice. In addition, the invention also relates to a blockchain technology, and the to-be-recognized voice can be stored in a blockchain. According to the method, training and prediction of the voiceprint recognition model are carried out by adopting the neural network, the voice features are mapped to the hidden spaces by adopting the self-attention mechanism to be processed respectively, and finally the processing results of all the audio frames are spliced to obtain the recognition result of the to-be-recognized voice, so that parallel operation can be realized, and the identification precision is considered while the identification efficiency is ensured.

The invention provides a method and device for identifying fraud and electronic equipment, and the method comprises the steps: identifying to-be-identified behavior data based on a classification model; when the to-be-identified behavior data is abnormal, determining an auditing result of the to-be-identified behavior data; taking the to-be-identified behavior data as a positive sample and addingthe positive sample into a preset sample set; taking the to-be-identified behavior data as historical behavior data, and extracting outlier historical behavior data corresponding to outliers in the historical behavior data based on a preset outlier detection model; and when the outlier historical behavior data belongs to fraudulent behaviors, adding the outlier historical behavior data as a positive sample into the sample set, and then training the classification model again based on the updated sample set. Through the technical scheme provided by the embodiment of the invention, the number ofsamples in the sample set is increased, the classification model can identify novel fraud behaviors in time, and the identification effect of the classification model is improved.

The invention relates to a curved surface object corner recognizing and spraying method. The method includes that on the basis of realization of fragmenting of a complex free curved surface, aiming at an included angle between adjacent surface fragment combinations, geometric relation with a spraying flare angle of a spray gun is adopted, and a law model of influence, of the included angles between the surface fragment combinations, on extreme thickness values of a coating during vertical spraying is built to recognize corner features. By the method, nonexistence of coatings and coating waste at junctions of concave surface fragments during vertical spraying at junctions of the surface fragment combinations can be effectively avoided, so that spraying efficiency is improved while spraying quality is guaranteed.

The embodiment of the invention discloses a training method, training equipment, a using method and using equipment of a multi-task recognition model, and a storage medium. An initial feature sharingmodel is trained through employing a sample image which comprises a plurality of face attributes with independent features, then a sample image which comprises a plurality of face attributes with coupled features is added for training, finally, a decoupling model part is added to the model, and a sample image containing a human face posture is added for training, so that the training of the multi-target task recognition model can be realized by gradually increasing human face attributes, and the sample image with the human face attributes with independent features can be fully utilized to realize feature sharing; and the decoupling effect on the coupling features is realized by adding the decoupling model, so that the recognition accuracy of the multi-task target recognition model can be effectively ensured, and the recognition efficiency can be ensured while the multi-face attribute recognition precision of the single model is ensured.

The invention provides an information recognition device, system and method, and relates to the technical field of information. The device comprises a scanning assembly and a disc type storage assembly, and the scanning assembly and the disc type storage assembly are connected with a server; the disc type storage assembly is located on one side of the scanning assembly. At least one container boxis stored in the disc type storage assembly, and each container box is provided with an electronic tag. The disc type storage assembly is used for rotating when receiving a rotation instruction from the server, so that the electronic tag on the container box in the disc type storage assembly is aligned with the scanning assembly; and the scanning assembly is used for scanning the electronic tag when receiving a scanning instruction from the server to obtain scanning information, and sending the scanning information to the server to enable the server to identify the scanning information. According to the invention, the accuracy of information recognition is effectively improved while the recognition efficiency is ensured.

The invention provides a method and a system for constructing an efficient quality inspection model of a self-learning and updating training sample, wherein two SqueezeNet Pro supervised learning models are adopted, fast k. The means algorithm acts as a preselected tag; The loss function of each model is optimized by comparing the two data enhancement algorithms. Conducting self-learning and training models; And update the weights of the parent model. The invention effectively improves the identification efficiency or the identification accuracy, improves the speed of the algorithm calculation, does not increase the time complexity of the operation, and can effectively prevent the occurrence of the local optimal situation.

The invention provides an industrial production data entity identification method based on Merkley-tree. The method comprises the following steps of: 1) for the floatability of industrial production data, performing corresponding standardized processing on the data by utilizing the properties of a matrix and a vector to ensure that numerical attribute values of the same entity are the same; 2) calculating the information entropy of each attribute column, obtaining attribute sensitivity information, removing attributes with low sensitivity, and sequencing the rest attributes in a descending order according to the sensitivity; 3) proposing a chain structure called St-Chain; performing progressive Hash coding based on the St-Chain , and dividing the entities with the same Hash value into thesame block; and 4) for the structure obtained in the step 3), continuing to calculate the hash value of the subsequent attribute in each tuple, and repeatedly dividing the tuples into blocks accordingto the difference of the hash values to finally obtain an entity identification result, thereby providing an entity identification method with moderate algorithm operation efficiency and high identification precision.

The invention provides an image content identification method and device, relates to the image identification technical field, and solves the image content identification problems in the machine vision field; the method comprises the following steps: receiving a to be indentified image; identifying the image according to an image identification model, wherein the image identification model comprises a first feature set having content information used for identifying image content; outputting an image identification result. The image content identification method and device can be applied to image content identification equipment having high accuracy requirements for image content identification.

Embodiments of the invention provide a method and a device for intercepting a multimedia message in a mobile device. The method comprises steps of extracting multimedia data from a multimedia message when it is monitored that the mobile device receives the multimedia message; determining the multimedia message is illegal or not according to multimedia data; and if yes, intercepting the multimedia message. Automatic adaption to change in illegal multimedia message is achieved, and the interception efficiency and the interception accuracy are improved.

The invention relates to the technical field of backplane bus communication, in particular to a communication system and method based on an SPI bus. According to the system and the method, a master station and at least one slave station connected with the master station in a daisy chain mode are adopted; the master station is used for sending a starting frame to the slave station; the slave station is used for generating and outputting a response frame corresponding to the starting frame after receiving the starting frame; the master station is further used for confirming the number of the slave stations based on all the response frames when the master station receives the starting frame and all the response frames, the communication system and method based on the SPI bus can achieve a slave station feedback mechanism, and the recognition efficiency and accuracy of the number of the connected slave stations are guaranteed.

The invention discloses a fingerprint image preprocessing method and device. The method includes the steps that a received fingerprint image is smoothed so that a first image can be obtained; the first image is processed through first enhancement so that a second image which is locally enhanced can be obtained; the second image is normalized so that a third image can be obtained; the fingerprint line direction angle and the fingerprint line frequency of the third image are calculated; the third image is processed through second enhancement so that a fourth image which is locally enhanced can be obtained; the fourth image is normalized so that a preprocessed output image can be obtained. According to the method and device, because the received fingerprint image is enhanced twice, the image quality of the fingerprint image enhanced twice can be high, the fidelity factor of the fingerprint line direction and the fingerprint line frequency of the preprocessed fingerprint image is high, and the identification accuracy and the identification efficiency of subsequent fingerprint recognition applications can be conveniently guaranteed.

The invention relates to a glass curtain wall crack detection method. The method comprises the following steps: step (1), acquiring a glass curtain wall image; (2) converting the glass curtain wall image into a gray level image; (3) calculating a first feature of the grayscale image based on an improved ULBP descriptor; step (4), carrying out dimension reduction on the first feature of the grayscale image to obtain a macro feature; and step (5), inputting the macro features into a classifier, and detecting whether cracks exist in the glass curtain wall image through the classifier. According to the invention, the crack of the glass curtain wall can be effectively detected, and the detection precision is high.

The invention provides a finger vein recognition method and system fusing basic features and deformation features. The recognition method mainly comprises the steps of image preprocessing, basic feature extraction, objective function construction and optimization and matching, deformation information extraction and basic feature and deformation feature fusion. According to the method, in the recognition process, deformation information is utilized, basic characteristics, matching scores and other information needed when the deformation characteristics are obtained are considered, and a fingervein recognition framework integrating the basic characteristics and the deformation characteristics is constructed. According to the method, the problem that when a single feature is used, the feature representation ability is insufficient is solved, the accuracy and robustness of finger vein recognition are improved, and the recognition efficiency is ensured.