38 results about "Digit recognition" patented technology

The invention provides an implementation method based on light-induced linear reconfigurable artificial synapses, and belongs to the technical field of neuromorphic electronic devices and intelligent sensing signal preprocessing. The method comprises the following steps: providing an artificial synapse transistor capable of regulating and controlling conductance through voltage pulses, irradiating visible light in a channel or a conductance modulation area of the artificial synapse transistor, applying continuous voltage pulses under an illumination condition, reading the conductance in a preset time window after the last pulse is ended, and taking the conductance as a weight value of a corresponding pulse sequence; by adjusting illumination parameters, conductance is converted from nonlinearity to linearity along with the pulse number relation, the nonlinear factor is reduced compared with the dark state, the linear interval is widened, and linearity reconstruction in the weight updating process is achieved. According to the method, reconfigurable linear weight modulation can be obtained only by means of a light field and simple pulse driving, the method is suitable for high-fidelity compression coding of 5-bit and above time sequence pulses, feature separability and recognition accuracy can be remarkably improved in handwritten number recognition, handwriting identification and other scenes, and training time and energy consumption are reduced.

The invention provides a multivariate data regulation and control processing method based on three-dimensional visualization, belongs to the technical field of image data processing, and aims at solving the problem that an existing AI intelligent inspection unit mainly depends on a two-dimensional image to carry out instrument detection and digital recognition. In order to overcome the defects of inaccurate state analysis, insufficient multi-view data fusion and the like caused by target occlusion misjudgment and spatial information loss in the prior art, a multivariate data system is constructed by fusing a three-dimensional point cloud, a depth map and a multi-view two-dimensional image, and three-dimensional reconstruction, spatial feature extraction and a multi-modal data fusion algorithm are combined, so that the multi-view data fusion is realized. Accurate detection of a power distribution station house equipment target, three-dimensional visual analysis of state parameters and dynamic regulation and control processing are realized. According to the method, the anti-shielding capability of target detection, the accuracy of state analysis and the real-time performance of data regulation and control are effectively improved, the intelligent level of operation and maintenance management of the power distribution station house is remarkably optimized, and the labor cost and the fault risk are reduced.

This invention discloses a license plate recognition method and system based on deep learning, comprising the following steps: acquiring license plate image data to generate an image dataset; performing semantic decomposition to divide the data into region code units, letter units, and number units, and constructing a semantic structure graph; constructing a multi-path character recognition model with an improved PCNN model as the backbone feature structure; inputting the image dataset into the multi-path character recognition model respectively, and outputting region code recognition results, letter recognition results, and number recognition results to generate candidate license plate recognition results; if there are illegal character combinations or missing structural unit information, performing path filtering processing to generate a candidate path set; performing inference imputation processing to generate a complete path set; and merging and outputting the license plate recognition results and the set of confident candidate paths. This invention improves the accuracy of character recognition while enhancing robustness to structural defects, possessing the advantages of high recognition accuracy and strong fault tolerance.

The present application relates to a kind of mixed STDP training method and system based on threshold increment cycle counting, construct including input layer, hidden layer and classification output layer STDP network architecture, and the threshold voltage increment of neuron is counted and controlled to the unsupervised learning process of STDP network architecture, the supervised learning process of combination classification output layer adjusts the weight value corresponding to neuron, by applying threshold increment cycle counting and neuron responsivity screening, effectively constrain weight update, reduce the weight update frequency, to greatly reduce the overall amount of computation of training, help to improve training learning efficiency.The present application method is applicable to handwritten numeral recognition, compared with traditional STDP training algorithm, can reduce 2.5 to 3.5 times amount of computation and improve about 3% classification accuracy.

The present application relates to the technical field of digital recognition, and particularly provides a nixie tube content recognition method and device, electronic equipment and storage medium, wherein the method comprises: acquiring a continuous video image sequence of a nixie tube, acquiring position information of a nixie tube character display segment according to a preset frame number by using a template matching algorithm; recognizing bright and dark states of the nixie tube character display segment by using a support vector machine (SVM); acquiring a brightness sequence unit of the nixie tube character display segment under the preset frame number according to a preset nixie tube arrangement mode; classifying the brightness sequence unit by using a deep learning method and outputting a recognition result; and the device comprises: a nixie tube and a PC terminal connected with the nixie tube, the nixie tube and the PC terminal are configured in an asynchronous state, and the PC terminal comprises: an acquisition module, a position recognition unit, a brightness recognition unit and a training module. The purpose is to solve the misrecognition problem caused by uneven brightness and defects in the nixie tube content recognition process.

The invention discloses a medical B-ultrasonic image automatic analysis system and method based on deep learning. The system comprises an image preprocessing module, a follicle detection module, a digital recognition module, an OCR text extraction module, an intelligent matching module and a report generation module. The method comprises the following steps: performing tilt correction, region extraction and perspective transformation on a B ultrasonic image, detecting follicle positions and morphological parameters by using a YOLO-11 model based on a rotating bounding box, and identifying digital labels; extracting an image bottom size text by adopting a CnOCR engine, and correcting to obtain structured data; intelligently matching follicles, numbers and sizes by calculating the Euclidean distance, and establishing triple association; and automatically generating a structured JSON report in combination with an anomaly detection mechanism. According to the method, full-process automation of follicle monitoring is achieved, the analysis efficiency, accuracy and consistency are remarkably improved, manual dependence is reduced, and reliable support is provided for clinical decision-making of assisted reproduction.

The application provides an image recognition simulation model and system, the image recognition simulation model comprising: an optical diffraction neural network for a handwritten digit recognition task, the optical diffraction neural network comprising: an input layer, a diffraction layer and an output layer; the input layer modulates incident light into a wavefront array carrying image information; the diffraction layer performs multiple diffractions on the light output by the input layer to simulate a convolution weight matrix, and adjusts a phase distribution in a manner of error back propagation; and the output layer generates a focused light spot corresponding to the number of categories, the focused light spot focusing on a number corresponding to a plurality of detection points, and a detection surface light spot intensity distribution of the focused light spot is used to represent a classification probability. Through nano structure design, high degree of freedom regulation of light field amplitude, phase and polarization is realized in a two-dimensional plane; meanwhile, through multi-layer cascading extension network depth, feature abstraction capability is improved, nonlinear mapping is realized through multi-layer cascading diffraction, and a classification light intensity distribution is directly output, and the light intensity distribution represents a classification result.

According to the hardware-network collaborative optimization FPGA digital recognition system and method, the LeNet-5 network architecture is improved, the number of channels is adjusted to be a multiple of 4 so as to be matched with an FPGA parallel computing array, a ReLU activation function is adopted to replace traditional Sigmoid / Tanh, and 8-bit integer quantization processing is implemented; a four-level assembly line architecture (convolution-quantization-pooling-activation) is adopted in the aspect of hardware, a snakelike movement strategy and a double-Block dynamic storage scheme are combined, the energy efficiency ratio and the reasoning speed are remarkably improved while high recognition precision is kept, an end-to-end real-time processing function chain from image collection, neural network reasoning to result display is achieved, and the method has the advantages of being high in recognition precision and high in practicability. The data reuse rate is obviously improved, the storage access bandwidth is reduced, a referenceable low-power-consumption and high-performance solution is provided for edge computing applications such as intelligent monitoring and industrial quality inspection, and high real-time performance and practicability are achieved.

The invention relates to the technical field of artificial synapses, in particular to a composite artificial synapse which comprises a core functional layer, the core functional layer is a composite material layer, and the composite material layer is formed by mixing a SrAl2O4-based long afterglow material and a NaYF4-based up-conversion material; eu < 2 + > and Dy < 3 + > are doped in the SrAl2O4-based long afterglow material; yb < 3 + > and Tm < 3 + > are doped in the NaYF4-based up-conversion material; under excitation of near-infrared light, the NaYF4-based up-conversion material transmits energy to the SrAl2O4-based long-afterglow material through up-conversion luminescence, so that the SrAl2O4-based long-afterglow material generates long-afterglow luminescence, and a synaptic function is realized; through thermal stimulation, the release process of defect trap electrons in the SrAl2O4-based long afterglow material is regulated and controlled, and thermally induced synaptic plasticity is realized. According to the invention, SrAl2O4: Eu < 2 + >, Dy < 3 + > and NaYF4: Yb < 3 + >, Tm < 3 + > are coupled, so that long afterglow luminescence (LPL) of near-infrared triggering and thermal regulation is realized, and information encryption, handwritten number recognition and mechanical arm gesture control are realized by artificial synapses based on photo-thermal synergistic stimulation.

A roller cutter for a raise boring tool includes a journal and a cutter shell positioned radially around the journal. The journal includes an external surface from which a journal end protrudes from. A seal retainer is positioned between the journal and the roller cutter, and wherein at least one digital identification tag is located on the roller cutter.

The invention relates to the technical field of artificial intelligence, and provides a handwritten numeral recognition method and related equipment based on multi-modal large model guidance, and the method comprises the steps: constructing a meta-training data set; performing style analysis on the support set image based on a multi-modal large model, generating a style description text, and encoding the style description text into a first prompt vector; performing off-line training on the small model according to the support set image and the first prompt vector based on a meta-learning framework; in an online recognition stage, generating a personalized style description text according to an unlabeled handwritten digital image of a new user through a multi-modal large model, and encoding the personalized style description text into a second prompt vector; adopting a self-supervised learning mechanism to carry out personalized adaptation processing on the trained meta-learning small model based on a second prompt vector; and recognizing a subsequently input handwritten digital image through the adapted meta-learning small model to generate a recognition result. According to the method and the device, high-precision recognition of personalized handwritten numbers can be quickly realized under the condition of no labeled sample, and meanwhile, the consumption of computing resources is reduced.

The application relates to a clock-drawing test scoring method and device, computer equipment and a storage medium. The method comprises the following steps: inputting a clock-drawing image dataset into an FCOS network to identify handwritten numerals in the clock-drawing image dataset; performing circle detection on the clock-drawing image dataset by using a maximum connected domain algorithm to obtain a clock region; based on the identification result of the handwritten numerals and the detection result of the clock region, performing pointer detection on the clock-drawing image dataset by using an image processing algorithm to obtain a pointer image of the clock-drawing image dataset, and performing hour hand and minute hand classification on the pointer image by using a breadth-first algorithm; scoring the handwritten numeral identification result, the clock region detection result, the pointer image detection result and the hour hand and minute hand classification result according to a set scoring rule, and adding all the scoring results to generate a clock-drawing test scoring result. The application can avoid evaluation fluctuations and differences caused by subjectivity and bias of artificial evaluation.

The invention belongs to the technical field of digital identification, and relates to a digital instrument test data auxiliary system, which comprises a data capture module, a data processing module, a data storage and judgment module and a human-computer interaction interface, the data capturing module collects the real-time state of an instrument data area through a monitoring camera, converts collected videos into pictures and transmits the pictures to the data processing module. The data processing module is used for carrying out region selection and image processing on the transmitted picture through an image processing system in computer software, and identifying digits in the picture; the data storage judgment module is used for transmitting the digits converted by the image processing system to a general data storage format or a database and marking the recording time of the digits; a user operates the human-computer interaction interface to realize various video digital capture requirements. According to the system, collected character images can be converted into digits to be stored in a computer, and test monitoring visually perceived by testers on site is replaced.

This invention belongs to the field of speech recognition technology, and discloses a digital speech recognition method, device, and medium, applicable to offline speech recognition modules on the edge and home appliances and consumer electronic devices with voice interaction functions. The method includes: acquiring the speech to be recognized; decoding the speech to be recognized through a speech recognition model to obtain a preliminary recognition result; reviewing N frames (N is a positive integer greater than 1) of phoneme probability output data to determine whether a preset digital confusion type exists; if it exists, correcting the preliminary result based on the corresponding confusion type's error correction strategy to obtain the final result; if it does not exist, directly outputting the preliminary result. This invention, through a process of "model decoding + frame review + targeted error correction," accurately solves problems such as path inclusion confusion and homonymous competition confusion in digital speech recognition. While ensuring real-time recognition, it significantly improves the accuracy of digital recognition, achieving low cost, high flexibility, and adaptability to the scenario requirements of edge devices.

The invention belongs to the technical field of artificial intelligence, and discloses a high-dimensional data dimension reduction system and method for semi-supervised local adaptive discriminant analysis, and the method comprises the following specific steps: 1), learning two independent projection matrixes for labeled data and unlabeled data; 2) setting residual constraints through the two projections to approach each other, and jointly converging to a unified low-dimensional space capable of distinguishing categories and reflecting an internal structure of data; (3) the projection of the labeled data can keep category discrimination information to the greatest extent, and the projection of the non-labeled data can keep the global manifold structure of the data optimally; the method is realized through a unified and solvable optimization objective function, and the method simultaneously utilizes marked and unmarked data, adaptively captures a local structure, is high in anti-noise capability, is suitable for semi-supervised high-dimensional data processing scenes such as face / handwritten digital recognition and gene data analysis, and shows remarkable advantages in classification on a standard data set.

The application discloses a kind of based on deep learning's digital identification method and system, method includes image acquisition, image optimization model design, digital number dimension flexible unit design, digital bit field navigation unit design, digital image recognition model establishment and digital identification.The application belongs to the field of image processing, specifically refers to a kind of based on deep learning's digital identification method and system, the scheme is through digital number dimension flexible unit, according to the dynamic adjustment receptive field of digital real size, based on digital bit field navigation unit, through forward control and reverse refinement, improve the accuracy of digital positioning;Based on contour weighted loss and morphological perception loss, strengthen the segmentation integrity of digital contour fuzzy area;Through double-way synchronous optimization, balance key stroke detail preservation and anti-interference ability;Based on the weight generated by digital pixel peripheral zone dispersity, and introduce structure difference loss, significantly improve the feature learning of digital contour area;Further improve digital identification effect.

The invention relates to a power distribution cabinet label and indicator light identification method based on deep learning and a related component, and is applied to the technical field of power distribution cabinet state identification. Target detection is carried out on a power distribution cabinet image through an improved Yov8 model, power distribution cabinet label coordinate information and indicator light information are determined, a feature extraction network of the improved Yov8 model adopts a multi-branch feature extraction structure, and a color perception feature enhancement module and a channel attention SE module are introduced; if the power distribution cabinet label coordinate information is not null, identifying the power distribution cabinet image based on an OCR digital identification model and the power distribution cabinet label coordinate information, and determining a power distribution cabinet label; based on the power distribution cabinet label coordinate information and the indicator lamp information, matching the power distribution cabinet label and the indicator lamp category, and determining the state of the power distribution cabinet. The method has the effect of accurately and efficiently identifying the state of the power distribution cabinet.

The present invention relates generally to the field of cybersecurity, and specifically to a system and method for registering, identifying and authenticating entities on digital platforms. The invention aims to address the shortcomings in secure and reliable digital identification and authentication across various digital platforms. The system comprises several integrated components: a registration and identification module for collecting and verifying entity information; an identifier generator or capturer for creating, assigning or capturing a unique identifier for registered entities; a comprehensive registration database for information storage; an authentication module for identity verification during digital access attempts; a digital footprint capturer for monitoring and recording online activities of entities; a notification module for issuing alerts about potential cyber threats; and a platform for supporting requests related to entity authentication, detailed information retrieval and suspicious cyber activity reporting. The invention significantly facilitates secure digital interactions by eliminating security loopholes in existing digital identification systems, thereby providing enhanced security and trust in online environments.

The application discloses a liquid level type instrument reading method and system, electronic equipment and a medium. The reading method comprises the following steps: performing image segmentation processing on a liquid level type instrument image according to vertical texture information of the instrument image, so as to determine a scale panel region and a liquid level column region in the instrument image; performing text recognition on the scale panel region, so as to obtain a scale number recognition result; performing image processing on the liquid level column region, so as to obtain a liquid level height; and determining a reading result of the liquid level type instrument according to the liquid level height and the scale number recognition result. There is strong vertical texture between the scale panel region and the liquid level column region. The image segmentation processing is performed on the instrument image according to the vertical texture information of the liquid level type instrument image. The segmentation line is determined in combination with the vertical texture information and the text detection result, the accuracy of segmentation is improved, and the accuracy of liquid level type instrument reading is improved.

This invention presents a rapid feature extraction method for railway infrastructure based on multimodal data fusion, addressing the issues of low efficiency and unstable accuracy in existing railway infrastructure identification methods. It falls under the field of intelligent railway detection and rail transit environment perception technology. The invention includes: acquiring 3D point cloud data from a LiDAR scanner along the railway line; converting the 3D point cloud data into a 2D depth image, where pixel values ​​represent laser ranging depth; extracting rod-shaped infrastructure using the depth image; statistically analyzing the reflection intensity from the 3D point cloud data to identify high-reflectivity point cloud areas as metallic infrastructure features; acquiring images of the railway line environment; extracting kilometer marker bounding boxes based on deep learning; and performing digital recognition on the image regions within the kilometer marker bounding boxes.

The present disclosure provides a multiplexed metasurface optical neural network device including metasurface layers that modify amplitude and phase of incident light to perform distinct classification tasks, each task associated with a degree of freedom including wavelength or polarization, and a detector that captures output intensity information encoding classification weights. A method includes projecting light intensity profiles representing objects onto metasurface layers using different wavelengths or polarization states, and capturing output intensity information corresponding to classification into different sets of classes. The device enables handwritten digit recognition at one wavelength or polarization state and object classification at another. The device also operates as a generative model, with metasurface layers generating diverse output images in response to random input light intensity profiles through spatial multiplexing, where random profiles are derived from a standard Gaussian distribution representing latent variables.

The present application relates to the technical field of computer vision, and proposes a kind of word wheel type water meter reading recognition method and device based on deep learning, comprising: obtaining word wheel type water meter image, and obtaining the digital area position image in the word wheel type water meter image by digital area detection model;The digital area position image is input into the network model based on rectification segmentation algorithm to obtain the crop image of target area by rectification and segmentation;The crop image of target area is input into digital recognition model after affine transformation, and a 10xQ matrix is generated as the output of model.The water meter reading detection model provided in the present application can recognize reading without character cutting, realizes the end-to-end recognition of word wheel type water meter reading, improves the accuracy of water meter reading recognition, in addition, the network model provided in the present application is very small in size, easy to deploy on mobile device, to provide more convenient way for water meter data reading and review.

The invention discloses a preschool child digital cognition operation box with an acousto-optic prompt, and relates to the technical field of preschool education, and the technical key points are that the preschool child digital cognition operation box comprises an operation box body, the upper end of the operation box body is hinged with a cover plate, the inner wall of the operation box body is provided with a cognition assembly, the front side of the operation box body is provided with a chute, and the front side of the operation box body is provided with an acousto-optic prompt. And the inner wall of the sliding groove is slidably connected with a storage box, and an acousto-optic prompt assembly is embedded in the operation box body. According to the invention, the number keys in the cognitive assembly correspond to the number of the placing grooves, and the number range covers 1-10, so that children's number recognition and number matching practice can be realized, preschool children are helped to quickly establish basic cognition of 1-10 digits, and the memory effect is enhanced; and meanwhile, the sleeve rod is matched with the sleeve ring, so that the children can be guided to complete corresponding exercises of numbers and numbers, and the digital cognition effect is further enhanced.

Use of a passive chip system for secure and contactless identification of persons, vehicles, goods and objects without the need for an active power supply or internet connection.

The invention discloses a picture digital recognition method based on a residual convolutional network, and relates to the technical field of digital recognition. Comprising the following steps: obtaining a to-be-identified picture, and preprocessing the to-be-identified picture to obtain a plurality of preprocessed segmentation blocks and corresponding position codes; inputting the preprocessed segmentation blocks into a trained picture digital recognition model for recognition, and obtaining digital information corresponding to the segmentation blocks; wherein the image digital recognition model recognizes the to-be-recognized image based on a residual convolutional network, and comprises an input layer module, a convolution calculation module, a residual calculation module, a flattening module, a full connection module and an output layer module; based on the recognized digits corresponding to the segmentation blocks and the position codes, carrying out combination to obtain a numeric string contained in the to-be-recognized picture; wherein the digits comprise printed digits and handwritten digits. And the efficiency and accuracy of number identification in the picture are improved.

This invention belongs to the fields of flexible electronic sensing, pattern recognition, and information security technology, and particularly relates to an integrated system and method for handwritten digit recognition and identity anti-counterfeiting based on a flexible pressure sensing array. It solves the technical problems of existing technologies, such as the inability to simultaneously capture dynamic pressure behavior features of handwritten digits, models that are unsuitable for small sample and low-resource scenarios, and the disconnect between content recognition and anti-counterfeiting verification. The invention achieves lightweight hardware, low-power algorithm, and integrated verification.

The application relates to a Cs3Bi2Br9-MoS2 heterojunction-based dual-terminal neuromorphic device. A zero-dimensional / two-dimensional hybrid structure is constructed through in-situ deposition, and charge accumulation injection and slow release are realized by using energy level offset and interface electron potential well, so that synaptic plasticity with a high pair pulse promotion index (PPF index > 230%), a high long-term potentiation curve linearity (R 2 > 0.97 and a photocurrent lag time > 800 s is exhibited. The device can realize 96% handwriting recognition accuracy in a convolutional neural network, and the accuracy is kept to be > 85% in a 13-75% RH wide humidity range, so that the device is suitable for neural network calculation in a complex environment.

The application discloses a kind of based on FPGA's pulse neural network handwritten numeral identification system, belong to integrated circuit field.It includes: CPU / GPU end software training module, parameter fixed-point quantization and export module, test input generation module, FPGA end model prediction module and simulation verification module.From snnTorch training to FPGA inference end-to-end code closed loop, make training parameter and RTL inference structure consistent;Through.coe / .mem double format export, make parameter initialization and simulation verification repeatable, can regress;Through the unified scheduling of TOP internal state machine "loading-replay-interval-complete" rhythm and interlayer parallel conversion, make two full connection and LIF, ROM IP core and LIF update strictly aligned, facilitate timing verification and parameter tuning;Through the generation of test input generation script with label and multi-image simulation test platform, realize stable regression test before board and fault location.

The invention discloses a digital ID recognition system for parts on an automatic assembly line, and relates to the technical field of digital recognition. Comprising a visual sensor module which is installed on a fixed support right above an automatic assembly line and used for shooting part shell pictures of moving parts on the assembly line in real time; wherein the part shell picture comprises a digital ID (Identity) of the part; the industrial personal computer is used for preprocessing the obtained part shell picture to obtain a plurality of preprocessed segmentation blocks and corresponding position codes, and inputting the preprocessed segmentation blocks into a trained picture digital recognition model for recognition to obtain digits corresponding to the segmentation blocks; and based on the identified numbers corresponding to the segmentation blocks and the corresponding position codes, carrying out combination to obtain the number ID of the component contained in the component shell picture. Component marks on the assembly line are accurately, efficiently and dynamically recognized, and automatic feeding is carried out.