111 results about "Optical character recognition" patented technology

The embodiment of the present application discloses a kind of plaque information acquisition method and device.The method and device of the present application according to the characteristics that plaque information is generally represented as attribute name plus attribute value key-value pair, adopt optical character recognition to extract plaque image text box, obtain the multiple different text boxes that are gathered in different regions on image, then, for text box classification, determine the category of text box, merge unit information text box with the attribute value text box closest to it, whereby, determine to obtain all attribute value text box and attribute name text box, again according to position information matching attribute value text box and attribute name text box, extract corresponding relationship, after matching is completed, the text information corresponding to is extracted to form the key-value pair of plaque information.Thereby, without complex recognition model, also without substantially increasing computational complexity, can accurately identify plaque information, facilitate offline staff to automatically collect plaque information.

Detecting and extracting contents from video data in text form is disclosed. Frames of the video data are analyzed to identify a scene, which includes a sequence of frames, that may include slide content. When the sequence of frames is determined to be slide content, optical character recognition is performed on one of the frames in the sequence to identify words and word coordinates. Post processing is then performed on the output of the optical character recognition to identify words, lines, objects, coordinates, and time codes. The output is stored as searchable text in a database. Video data can be searched based on text extracted from speech and / or text extracted from images.

Techniques are disclosed relating to extracting data from a document, using a large language model (LLM), to populate fields in a data structure. A computer system may receive a request to populate multiple fields of a data structure with data extracted from text of a document. The computer system parses the text using an LLM (as well as regular expressions or other parsing techniques in some embodiments). The parsing includes issuing, to the LLM, a sequence of queries targeting individual ones of the multiple fields. The computer system applies a validation algorithm to results received from the LLM in response to the sequence of queries. The validation algorithm confirms the presence of results in the text of the document and populates the data structured with the validated results. In various embodiments, the computer system performs an optical character recognition (OCR) on the document to determine the text for parsing.

The present application relates to the field of character recognition and image processing, and particularly relates to an ancient book and handwritten character intelligent recognition method and system. The method comprises: obtaining a whole page image by scanning an original page of an ancient book document, sequentially performing correction, enhancement, denoising and standardization processing to obtain an input image and a text label; using a character detection model to segment a character region and generate a text box coordinate, classifying a handwritten and printed character block, and establishing a corresponding relationship between a region, a coordinate, a category and a text label; performing horizontal and vertical projection to generate a projection block, calculating a sliding window parameter through a connected domain and a reference height, and adjusting a boundary to generate a recognition block; performing unified character processing and character recognition to obtain a result, triggering dynamic resolution optical character recognition and boundary adjustment through joint judgment, and outputting an overall detection and recognition result and a returnable recognition block. The present application improves the recognition accuracy and adaptability through closed-loop reprocessing and dynamic adjustment of the recognition block.

The application provides an optical character recognition method and system, comprising: pre-training a detection error correction hybrid model together with a text recognition model to be trained to obtain a trained text recognition model; inputting a text image to be recognized into the trained text recognition model to obtain a text recognition result output by the text recognition model. The application considers the performance problem in the actual operation process of the model, the detection error correction hybrid model does not participate in the calculation in the actual operation stage, and only participates in the pre-training in the process of pre-training the text recognition model to improve the text recognition result of the text recognition model, and then correct the model parameters of the text recognition model to improve the overall recognition performance of the text recognition model, solving the defect that the text recognition is not accurate in the prior art, effectively improving the recognition accuracy without affecting the recognition speed of the text recognition model.

This application relates to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for processing wiring diagrams of substations. The method includes: acquiring an image of a substation wiring diagram and performing quality enhancement processing on the image to obtain a processed image; performing optical character recognition (OCR) to obtain a title recognition result and determine the diagram type; based on the diagram type, performing region segmentation on the processed image to identify and locate the wiring diagram area, text area, and table area; parsing to obtain the structural data of each device in the wiring diagram; converting the data into standardized data for the power industry; and, provided the standardized data passes rule review, performing similarity matching between the standardized data and reference standardized data to obtain a matching result; and based on the matching result, outputting the quality review result of the wiring diagram. This method can improve the accuracy of wiring diagram review.

This application discloses a method and apparatus for calligraphy and painting recognition based on deep learning, belonging to the field of artificial intelligence and digital image processing technology. The method includes: acquiring images of calligraphy and painting works; preprocessing and enhancing restoration by introducing a generative adversarial network with an edge-preserving loss function to preserve the texture of the ink marks; extracting the character structure features of the signature, inscription, and seal areas; inputting them into a specially trained optical character recognition and semantic analysis hybrid model; during decoding, weighted fusion of visual feature vectors and contextual semantic probability vectors to output initial recognition text; combining a calligraphy dictionary database, performing semantic error correction and matching through a joint calculation formula of quantified character topological edit distance and semantic probability to obtain and output the target recognition text. This solves the problems of failure due to lack of topological repair capability for damaged pixels and inability to hit the correct characters, achieving high-precision recognition and deep knowledge interpretation of calligraphy and painting text.

Techniques are disclosed relating to extracting data from a document, using a large language model (LLM), to populate fields in a data structure. A computer system may receive a request to populate multiple fields of a data structure with data extracted from text of a document. The computer system parses the text using an LLM (as well as regular expressions or other parsing techniques in some embodiments). The parsing includes issuing, to the LLM, a sequence of queries targeting individual ones of the multiple fields. The computer system applies a validation algorithm to results received from the LLM in response to the sequence of queries. The validation algorithm confirms the presence of results in the text of the document and populates the data structured with the validated results. In various embodiments, the computer system performs an optical character recognition (OCR) on the document to determine the text for parsing.

Provide a system. 【Solution means】 Means for receiving image information from a communication device, Optical character recognition means for extracting character information from the received image information, Automatic journalizing means for classifying the extracted character information into accounting items, Means for generating a financial report based on the classified information, Means for providing advice based on the generated financial report and classified information, Means for managing the interaction with the user and analyzing change requests from the user, Means for integrating and managing daily settlement information based on the image information and visualizing the user's financial status, A system including the above.

The application discloses a text recognition method, belongs to the technical field of optical character recognition, and helps to improve text recognition accuracy. The method comprises the following steps: inputting a target image into a convolutional neural network in a pre-trained character recognition model, obtaining a feature map with a height of D and a width of n output by the convolutional neural network for the target image, wherein D and n are integers greater than 1; reorganizing the feature map to obtain a feature sequence of the target image; encoding and mapping the feature sequence through a sequence encoding network in the character recognition model to obtain an encoded sequence; and decoding the encoded sequence through a CTC decoder in the character recognition model to obtain a character recognition result of the target image. According to the method, the feature map with a height greater than 1 is extracted, so that character recognition can be performed based on more fine-grained features, and the text recognition accuracy of complex texts such as arc-shaped text images and seal images is improved.

The application discloses a calendar filling method, system and device based on optical character recognition, which obtains a calendar panel picture and a target date to be filled in a webpage, configures a page turning button for turning a text area page on the calendar panel picture, intercepts a text area from the calendar panel picture, identifies the year and month in the text area through an optical character recognition technology, adjusts the year and month in the text area through the page turning button, so that the year and month in the text area are consistent with the year and month in the target date, identifies all the days in the text area through the optical character recognition technology, calculates the position of the day in the target date in the text area according to all the days in the text area and the day in the target date, and selects the position to complete the filling of the target date in the webpage. The application realizes modularization, is compatible with all calendar filling modes, and saves development time cost.

In an approach, a processor performs document layout analysis on a document generating a plurality of textual regions; extracts characteristics from each of the plurality of textual regions and associates the respective characteristics to the respective textual region as metadata; classifies each of the plurality of textual regions as an optical character recognition (OCR) region, non-OCR valuable region, or non-OCR non-valuable region using a classifier; performs OCR on each OCR region generating an OCR output; identifies associated constant OCR data from a constant OCR data repository for each non-OCR valuable region; merges the associated constant OCR data with the OCR output generating a complete OCR data for the received document; performs data extraction on the complete OCR data to identify data fields and key-value pairs generating extracted data; and determines whether the extracted data is valid based on a set of rules.

The application discloses a secondary terminal number identification method and device and a medium, relates to the secondary circuit operation and maintenance technical field, and comprises the following steps: acquiring a terminal row image of a target area; determining the number text area of each terminal based on deep learning according to the terminal row image; performing optical character recognition on each number text area to obtain the original identification number of each terminal; adaptively correcting the original identification number of each terminal according to the coding rule and context information to obtain each corrected number; associating each corrected number with a preset engineering file, completing the connection information of each terminal, and obtaining the topological relationship between the terminals and devices in the target area; the preset engineering file stores the connection relationship between the terminals and devices; and the topological relationship is visually displayed, so that the secondary terminal operation and maintenance efficiency can be improved.

The application relates to the technical field of artificial intelligence optical character recognition, and provides an image character recognition method, device, computer equipment and storage medium, the method comprises the following steps: respectively performing character recognition and semantic segmentation on a to-be-recognized image to obtain text information and a Mask image; determining the space width and the pixel height sequence of the text according to the text information and the Mask image, and obtaining space position information according to the space width and the pixel height sequence; and obtaining an image character recognition result according to the text information and the space position information. The method can improve the accuracy of recognition in multiple scenes.

This invention relates to the field of advertising screen content detection technology, and particularly to a method for detecting and identifying illegal content on advertising screens based on a depth graph convolutional network. The method includes: segmenting the display area of ​​the advertising screen playback interface image; performing optical character recognition (OCR) processing on the text display area image and extracting features from the image display area image; constructing a heterogeneous relationship graph based on text node feature vectors and visual node feature vectors; inputting the heterogeneous relationship graph into a depth graph convolutional network to propagate and aggregate the text node feature vectors and visual node feature vectors; performing global graph pooling processing on the updated text node feature vectors and updated visual node feature vectors; and inputting the global graph representation vector into a classification layer for illegal content category determination. This invention can jointly analyze the text and image content simultaneously contained in the advertising screen playback interface, improving the accuracy of illegal content identification.

A system for converting digital documents into American Disabilities Act (ADA) Web Content Accessibility Guidelines (WCAG) compliant content includes a content upload system capable of receiving a digital document. An optical character recognition program converts the digital document to a plain text document. A language module structurally organizes the plain text document while maintaining the content of the digital document. A hypertext markup language (HTML) module builds an HTML based document having a structure that is WCAG compliant.

This invention provides a method and apparatus for recognizing customs declaration documents, relating to the field of artificial intelligence technology. The method includes: acquiring a target image of a customs declaration document to be recognized; inputting the target image and prompt words into a visual language model to obtain a first recognition result output by the visual language model; the prompt words are generated based on preset business requirements; the visual language model is obtained by performing low-rank adaptive fine-tuning of a CogVLM40B model; performing optical character recognition on the target image to obtain a second recognition result; and correcting the first recognition result based on the second recognition result to obtain a target recognition result corresponding to the customs declaration document. The customs declaration document recognition method provided by this invention improves the efficiency and accuracy of customs declaration document recognition by integrating the semantic understanding advantages of visual language models and the character accuracy advantages of optical character recognition through automated collaborative verification.

The application relates to a method and device for judging the validity of character information and an electronic device. The method comprises: obtaining a character name to be judged and a corresponding video screenshot, wherein the character name to be judged is extracted from the video screenshot by optical character recognition technology or visual recognition technology; constructing a prompt word according to the character name to be judged, wherein the prompt word comprises a role definition, a text to be judged, a judgment task, a constraint condition and an output requirement; inputting the prompt word and the video screenshot into a target visual large model to enable the target visual large model to perform image understanding, semantic understanding and joint reasoning on the video screenshot and the character name to be judged according to the prompt word, and generate an output result comprising a judgment result and a judgment reason; and analyzing the output result to determine the validity of the character name to be judged and marking the character name to be judged. The application solves the technical problem that the existing method cannot accurately judge the validity of the character name in a video.

The application discloses a file automatic stamping method, device, medium and equipment. The method comprises the following steps: acquiring a file image to be processed, and preprocessing the file image to obtain a first processing result; performing optical character recognition and semantic analysis on the first processing result, extracting key text information of the file, and obtaining a second processing result; judging whether stamping is needed and determining a target area of an electronic seal based on a preset stamping rule and the second processing result, and obtaining a third processing result; and synthesizing a specified electronic seal in the target area according to the third processing result, and generating a stamped file.

Methods and systems for remote check deposit are disclosed. A check for deposit is processed without the need for a server to receive any image of the check initially. Instead, optical character recognition (OCR) data is received at the server from a mobile device. Verification processing for the check is then performed using the OCR data. If the verification process is successful, a confirmation notification is sent to the mobile device. Subsequently, after sending the confirmation notification, a check image is received, from which the OCR data was determined. The check is, in turn, processed for deposit using the received check image.

Embodiments of the invention are directed to systems, methods, and computer program products for enhanced accuracy of image data obtained through optical character recognition (OCR) processing. In some embodiments, the method includes receiving a first image file obtained by an image capture device; performing OCR on the first image file to generate an OCR data file; inputting the OCR data file into at least two machine learning engines, where the at least two machine learning engines are configured to operate in parallel; based on an output of the at least two machine learning engines, generating a confidence score associated with the OCR data file; and based on the confidence score, processing the OCR data file over a real-time settlement rail.

The invention relates to a method for testing a computer program with a graphical user interface, wherein a test environment is provided for testing the system behavior of the computer program under test by operating at least one control element of the graphical user interface, wherein, in addition to the first image data stream for the display device, a second image data stream is provided to the test environment via the display interface of the data processing device, wherein, by means of optical character recognition, control elements operable by an input device in the graphical user interface provided via the second image data stream are recognized as an object of the graphical user interface.wherein at least the operable controls recognized in the graphical user interface are entered as input data into a large language model connected to the test environment via a data interface, with the stipulation that an action option for operating at least one control element of the graphical user interface is obtained at the data interface, wherein an input signal is generated by the test environment depending on the action option provided at the data interface for operating at least one recognized control element, and wherein the generated input signal is provided at the input interface of the data processing system in order to trigger a simulated user input for testing the computer program.

This invention relates to a method for diagnosing error sources in the manufacturing process of precision micro-connector molds. Addressing the challenge of integrating multi-source heterogeneous detection data and attributing errors during the manufacturing process, it proposes a method based on semantic fingerprint extraction and incremental fusion of knowledge graphs driven by multi-source raw data. Through industrial Ethernet, dedicated communication protocols, and optical character recognition, data from CNC machining, coordinate measuring machine (CMM), electrical discharge machining (EDM), and maintenance fault logs are collected in real time, and data feature compression, geometric fingerprinting, material response analysis, and semantic phrase standardization are performed. The core solution automatically constructs and evolves a knowledge graph through low-dimensional semantic fingerprint spatial mapping and anchor point expansion mechanisms, enabling real-time location of manufacturing error sources and the formation of causal chain attribution reasoning. This method achieves dynamic identification and ontology fusion of novel error patterns, improving the efficiency of anomaly diagnosis and the self-evolutionary capability of knowledge in the mold manufacturing process.

This application discloses a method, apparatus, device, and storage medium for setting up a tilting camera, including: determining a target imaging angle based on a preset shooting distance, a preset tilt angle, and an imaging angle algorithm; the preset shooting distance includes a near-end shooting distance and a far-end shooting distance; calculating the near-end imaging distance and the far-end imaging distance based on the near-end shooting distance, the far-end shooting distance, the target imaging angle, the preset tilt angle, and a preset imaging formula; determining a near-end imaging point on the camera's optical axis based on the near-end imaging distance, and determining a far-end imaging point based on the far-end imaging distance, the near-end imaging point, and the target imaging angle; determining the imaging image composed of the near-end imaging point and the far-end imaging point, and the target focal length corresponding to the imaging image; and setting the camera's focal length to the target focal length for image capture. This achieves simultaneous clarity of both near-end and far-end text in the image, reduces the requirements for camera depth of field, and improves the accuracy of optical character recognition.

Disclosed is a computer-implemented method that includes receiving a digital image or a scanned page of a historical lab notebook that contains handwritten text and a chemical structure drawing. The method includes performing optical character recognition on the image or the page to convert handwritten text into machine-readable text data. The method includes performing optical chemical structure recognition on the image or the page to identify hand-drawn chemical structures and reaction diagrams. The method includes translating each chemical structure and reaction diagram into a standardized digital representation. The method includes analyzing the recognized text data with a natural language processing engine to extract scientific context and metadata. The metadata includes identifying a chemical entity, a reaction condition, an experimental parameter, or a result. The method includes correlating the output of the chemical structure recognition and the natural language processing by associating each chemical structure with its textual context.

This application relates to the field of data processing technology, applied in fintech and smart healthcare scenarios. It provides a data processing method, apparatus, electronic device, and computer-readable storage medium. The method includes: acquiring an image to be recognized; performing optical character recognition processing on the image to be recognized to obtain preliminary character recognition results; formatting the preliminary character recognition results to obtain formatted recognition information; performing field extraction processing on the formatted recognition information based on a preset large language model to obtain field extraction results; if the field extraction results indicate abnormal field extraction, performing recognition processing on the image to be recognized based on a preset large visual model to obtain character recognition results; and adjusting the character recognition results based on a preset strategy to obtain adjusted recognition results. Through the above technical solution, the efficiency and accuracy of information extraction can be improved.

This invention discloses a qualification verification method and system based on intelligent parsing of structured bidding documents. The method includes: acquiring structured bidding documents and extracting qualification fields and their source attributes to be verified; calculating the field confidence score of each qualification field to be verified based on a field quality assessment model and comparing it with a corresponding preset verification trigger threshold; if the field confidence score is greater than or equal to the threshold, the field is used as a qualified input field, and the matching target verification interface is called to verify the authenticity of the qualification; if the field confidence score is less than the threshold, it is considered a low-quality field, and a field enhancement processing instruction is generated to perform image super-resolution reconstruction or optical character recognition secondary correction. By calculating the field confidence score through a field quality assessment model and performing verification or enhancement processing according to the threshold comparison results, the intelligence level and resource utilization efficiency of the verification process are improved, solving the problems of resource waste and inability to process low-quality fields caused by using incorrect fields to call the verification interface.

This specification provides an adaptive acquisition, parsing, and standardized archiving method and device for ophthalmic multimodal data, relating to the fields of medical internet and medical big data processing technology. The method includes: an access adaptation step: connecting to the physical interface of the target ophthalmic imaging device via a smart adapter terminal supporting multiple physical interfaces; a data acquisition step: establishing a current acquisition task in response to the current user's identity recognition operation; loading a corresponding protocol parsing plugin based on the feature information of the target ophthalmic imaging device to obtain the original image data captured by the device on the current user; a different device parameter parsing step: determining a parsing strategy for optical character recognition of the character information burned into the original image data based on the feature information of the target ophthalmic imaging device, and executing the parsing strategy to extract examination parameters and / or user information; verifying and standardizing the extracted examination parameters and / or user information with the target information of the current acquisition task.

This invention discloses a method for recording meeting content using handwritten annotations and multimodal fusion. The method includes the following steps: Multimodal trigger acquisition: Image capture of the target projection application's window client area; calculation of the change ratio based on inter-frame difference; triggering optical character recognition (OCR), whole-image multimodal understanding, and nearest neighbor speech recognition when the change ratio reaches a specified threshold; Handwritten input vectorization: Acquiring and rendering the user's handwritten input in the summary window, vectorizing the input into strokes and shapes; Handwriting semantic parsing and binding: Parsing the semantic category of the handwriting based on a handwritten grammar dictionary and context, and establishing anchor points for binding with summary items or chart areas; Multimodal content fusion: Fusing handwriting semantics with AI summarization. This invention reduces costs, and the high signal-to-noise ratio input makes the basic data for subsequent AI processing more accurate, laying a solid foundation for the accuracy and reliability of meeting recordings.