63 results about "Text detection" patented technology

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 discloses an identity card offline intelligent identification and correction system for public security actual combat, relates to the technical field of image processing and artificial intelligence, and comprises an image input module, an image correction module, an OCR identification engine module, an intelligent field extraction module, a backend reasoning core, a front-end interaction interface, a standardized output module and a local storage unit, and all calculations are completed on a local terminal and are independent of a network; the overall execution process is as follows: S1, image input; S2, certificate detection and correction; S3, text detection; S4, character recognition; S5, field extraction; and S6, result output. The identity card offline intelligent identification and correction system for public security actual combat adopts a full offline localization architecture, all image acquisition, model reasoning, data identification and result output are completed on a local terminal, are independent of a network and do not upload to a cloud, the risk of public security sensitive information leakage is avoided from the root, and the data security and confidentiality are significantly improved.

The application relates to a text motion effect video generation method and device, a storage medium and electronic equipment. The method comprises the following steps: performing a text detection operation on a target image to identify feature information of each text region in the target image, wherein the feature information comprises text content and boundary box coordinate information; according to the feature information of each text region and image information of the target image, filtering out a motion effect type corresponding to each text region from a preset motion effect library to obtain a motion effect template; generating a foreground image and a background image of the target image by using an image repair and completion model; and inputting the foreground image and the background image into an image generation model, so that the image generation model generates a text motion effect video of the target image according to time sequence parameters of the motion effect template. The application solves the technical problems of insufficient automation, low efficiency and low output quality of existing text motion effect generation.

Embodiments of the present application relate to the technical field of image processing, and disclose a scene text detection method based on polynomial fitting, which comprises the following steps: performing affine transformation on a text region frame marked on a sample image, so that the center of the text region frame after transformation is located at the origin and the line connecting the first and last points of the upper long side is in a horizontal state; performing polynomial fitting on the text region frame after transformation to obtain polynomial coefficients, and marking the polynomial coefficients as training labels; inputting the sample image into an initial network to obtain predicted polynomial coefficients, and iteratively training the initial network based on the predicted polynomial coefficients and the training labels until convergence, so as to obtain a trained detection model; and inputting an image to be detected into the detection model to output a text region in the image to be detected. The method needs to learn fewer parameters and all the parameters conform to visual rules, the pressure of model training is relatively small, and the trained detection model can accurately recognize text regions of any shape.

This invention discloses a text detection method and system based on cross-level feature enhancement and auxiliary feature guidance, belonging to the field of computer vision technology. The method includes the following steps: extracting feature maps of different levels from the image to be detected; for low-level features, dynamically adjusting the receptive field by adjusting the convolutional kernel size to obtain multi-scale information, generating enhanced features for low-level features; for high-level features, using differential convolution to enhance high-frequency information at text edges, generating enhanced features for high-level features; fusing features from different levels using cross-level feature enhancement to generate fused features; extracting semantic information from high-level features to generate auxiliary features; for the generated auxiliary features and fused features, using attention guidance to determine the attribution of text kernel pixels, obtaining attention-guided features; and performing text-to-non-text prediction on pixels to generate text bounding boxes. The scheme in this invention enables the network to more accurately segment text regions.

This application discloses a character recognition method. The method includes: performing text detection on a target image to obtain at least one polygonal region; determining a target polygonal region from the at least one polygonal region, wherein the target polygonal region has an irregular shape; determining four target vertices from multiple vertices of the target polygonal region, and using these four target vertices as the four vertices of a rectangular region corresponding to the target polygonal region; mapping pixel information of each pixel in the target polygonal region to the rectangular region based on the positions of each vertex of the rectangular region in the target image to generate pixel information of each pixel in the rectangular region; and performing OCR processing on the rectangular region and the polygonal regions in the at least one polygonal region other than the target polygonal region to obtain the character recognition result of the target image. This application embodiment can simultaneously achieve both accuracy and efficiency in character recognition.

This invention relates to the field of text detection, specifically to a text detection method, system, and storage medium in low-light environments. The method includes: acquiring a low-light text image to be detected, and performing adaptive image quality preprocessing on the low-light text image to obtain an enhanced text image; inputting the enhanced text image into a pre-trained target detection network to obtain the location coordinates of the text region in the text image; cropping the text region image based on the location coordinates of the text region, and inputting it into a pre-trained pixel segmentation network to obtain the text information contour in the text region image; extracting text information based on the text information contour, and inputting it into a pre-trained text recognition network for feature extraction and sequence prediction to form a target text sequence.

This invention proposes a text detection method generated by a large language model, comprising: segmenting the input text into multiple basic text units; extracting the rhetorical relationship types between the multiple basic text units and constructing a multi-relationship graph; extracting the first semantic expression features of the text segment corresponding to the first node and the second semantic expression features corresponding to the second node through a pre-trained language model; inputting the multi-relationship graph into a multi-relationship graph neural network model for feature learning, updating the first and second semantic expression features; and performing graph readout processing on the learned and updated first and second semantic expression features to generate probability distributions of different fine-grained categories corresponding to the input text. This method achieves fine-grained text detection.

This application provides a text steganography analysis method, apparatus, device, medium, and program product, which can be applied to the field of text detection. The method includes: in response to receiving text to be detected, inputting a first feature vector of the text to be detected into a semantic baseline extraction layer of an analysis model, extracting low-frequency semantic features and artifact dimension signal features from the first feature vector, and fusing the two to obtain a second feature vector; inputting the first feature vector into an artifact feature amplification layer to amplify the artifact dimension signal features in the first feature vector, obtaining a third feature vector including the amplified artifact dimension signal features; inputting the second feature vector and the third feature vector into a global representation extraction layer for feature fusion to obtain a fourth feature vector; inputting the fourth feature vector into a classification head module to obtain a detection result, the detection result including whether the text to be detected is normal text or steganographic text.

The application relates to the technical field of natural language processing, in particular to an AI generated text detection method and system based on a deep learning model. The method comprises the following steps: performing character normalization and sentence segmentation on a to-be-detected text, segmenting the text by a sliding window to generate a segmented sequence; constructing deep semantic discrimination features, logarithmic probability difference trace features, style structure features and entity consistency features; inputting a gate fusion discrimination network, outputting a generated probability and confidence degree after probability calibration, and outputting a segmented contribution degree; performing threshold adaptive determination according to a risk level, a text length and the confidence degree, outputting a review state and generating a traceable record when the confidence degree is insufficient; and triggering incremental updating based on drift monitoring and a feedback sample pool. The application improves cross-domain robustness and interpretability, reduces false positives and false negatives, and supports long-term stable operation.

This application relates to a method, device, storage medium, and computer program product for identifying the authenticity of returned or exchanged images. The method includes: acquiring text detection boxes and text recognition results for each segment of text in the returned or exchanged image; determining global and local information of the text in the returned or exchanged image based on the text detection boxes and text recognition results; determining a first determination result for the returned or exchanged image based on the global information, local information, and interface feature information corresponding to the target platform category; acquiring a screenshot of a region in the returned or exchanged image; inputting the screenshot and the target platform category into a preset classification network to obtain a second determination result regarding whether the screenshot belongs to a forged image; and determining whether the returned or exchanged image is a forged image based on the first and second determination results. This method helps improve the accuracy of identifying the authenticity of returned or exchanged images.

This invention presents an AI-generated text detection method based on graph structure features, belonging to the fields of artificial intelligence and natural language processing. The method includes: dataset construction, entity relation extraction and graph structure construction, graph structure feature extraction, graph structure feature model training, and text detection. It further incorporates traditional text feature extraction and model training, adaptively fusing the traditional text feature model and the graph feature model based on confidence-weighted entropy, and then performing text detection based on the fused model. This invention is the first to perform AI text detection from the perspective of graph structure features, breaking through the limitations of existing research that focuses on surface features such as vocabulary, syntax, and perplexity. The fusion strategy dynamically adjusts the fusion weights by quantifying the uncertainty of model predictions, maintaining a high level of performance on both original data and adversarial examples, achieving a balance between detection accuracy and adversarial robustness. It can be widely applied to the detection of AI-generated content such as news content and academic papers.

The application provides a box recognition method, system and device based on multi-view vision and a storage medium, and the method comprises the following steps: obtaining a first image and a second image at different angles; inputting the first image and the second image into an image recognition neural network respectively to obtain local pattern information, corner point position, parallax information and 3D matching information representing a box; inputting the local pattern information and the corner point position into a context adjustment layer of a character detection neural network to obtain stereo information corresponding to the local pattern representing the box; and obtaining a spatial box with relative positions based on the corner point position and the stereo information. The application can enhance the algorithm recognition effect, has good real-time effect, consumes less computing resources, is convenient to deploy on a resource-limited vehicle-mounted edge device, is not prone to causing misrecognition and collision, and improves driving safety.

This invention relates to a method and system for text detection and recognition in complex layout images. The method includes: performing text detection on a high-resolution original image and generating multiple candidate text boxes containing text; post-processing the multiple candidate text boxes based on the layout structure characteristics of the text to obtain multiple effective text regions; cropping the original image according to the effective text regions to obtain text sub-images corresponding to the effective text regions; and performing text recognition on each of the multiple text sub-images to obtain text recognition results. The method and system of this invention improve the accuracy, stability, and efficiency of text recognition in complex layout image scenarios.

The application discloses a morphing text detection method and device and a storage medium. The method comprises the following steps: extracting image features of a to-be-processed image, the to-be-processed image being an image collected from a specified text carrier, and the specified text carrier being easy to deform under stress; inputting the image features into a control point detection network to obtain an edge control point sequence; inputting the edge control point sequence into a differentiable B-spline layer to perform matrix multiplication calculation on a base function matrix and the edge control point sequence, thereby obtaining a B-spline curve point sequence, and the base function matrix being a two-dimensional matrix composed of B-spline base function values; and generating a text detection frame based on the B-spline curve point sequence. The operation of the differentiable B-spline layer is differentiable in a deep learning framework, and the gradient can be back propagated to the parameters of the control point detection network through the differentiable B-spline layer, thereby realizing integrated end-to-end optimization of text detection and curve fitting and significantly improving the performance of the model in a complex morphing text detection task.

This invention discloses a paragraph-level generated text detection method based on deep maximum mean difference (MDD) metric. The method includes: modeling generated text detection as a nonparametric two-sample test problem; designing a deep composite kernel to fuse shallow syntactic features and deep semantic features; training a deep kernel network with the goal of maximizing unbiased estimation of test power; in the testing phase, introducing a Wild Bootstrap mechanism to generate random weight sequences with autoregressive structures; performing a weighted quadratic transformation on the MMD kernel difference matrix; preserving paragraph sentence sequence correlation; accurately estimating the null distribution of the test statistic; and completing the detection by comparing the observed statistic with the test threshold. This method improves sensitivity to high-dimensional semantic differences, alleviates the problem of decreased test power caused by sequence correlation, and can efficiently distinguish between human and AI-generated text. It is suitable for generated text detection in scenarios such as news and academic writing.

This invention discloses a method for detecting and recognizing tilted and distorted text in natural scenes, comprising the following steps: preprocessing a natural scene image containing tilted, curved, or perspective-distorted text, and obtaining the stroke width information of each pixel in the image and the corresponding edge pixel relationship through edge detection and stroke width calculation. Based on the consistency of stroke width and the edge correspondence, connected region aggregation is performed on the pixels in the image to obtain candidate text regions, and a text stroke structure is constructed. The stroke center lines are further extracted and control points are generated. The weights of the control points are determined based on the stroke structure information, and a weighted control point set is constructed. Based on this control point set and candidate regions, a deformation mapping is generated using the moving least squares method to correct the tilted and distorted text. The corrected text regions are then filtered and character recognition is performed, outputting the text detection and recognition results. This invention is applicable to text processing in complex backgrounds.

This invention discloses an AIGC text detection method and system based on multi-dimensional feature fusion, belonging to the field of natural language processing. It addresses the performance bottlenecks and interpretability issues of existing technologies in text rewriting / polishing and multi-model hybrid content detection. Closed-loop detection is achieved through S1 multi-dimensional feature extraction (four orthogonal dimensions: statistical linguistics, semantics and fluency, syntactic structure, and information theory), S2 attention-weighted feature fusion (Z-score normalization + dynamic weighting by an attention network), S3 adaptive threshold determination and confidence evaluation (dynamically generating thresholds based on text metadata), and S4 interpretability report generation. The system includes feature extraction, fusion, discrimination, and report generation modules, as well as offline construction and online inference units. It balances detection accuracy, anti-circumvention robustness, cross-scenario generalization ability, and result interpretability, with low inference costs and easy scalability.

The application discloses a geological large model construction method and system based on video processing, comprising: acquiring a geological data set; using a U-Net of lithology filling area semantic segmentation to perform visual analysis on geological maps to obtain a lithology filling area mask; using a text detection network and a text recognition network cascade model to perform OCR and field structured extraction on scanned documents and map them to a geological data dictionary; using a space-time Transformer to perform space-time feature coding on a deformation sequence of a remote sensing and InSAR time sequence video to generate a space-time embedding representation; constructing a cross-modal sample unit and aligning it to a unified geological object primary key index to form a traceable training sample; and based on the sample, training a geological basic large model containing a video encoder to generate a target task result through cross-modal fusion and a task output head. The application improves multi-source alignment efficiency and reasoning reviewability, and greatly improves the accuracy of geological data processing.

This application discloses an image data processing method, apparatus, storage medium, and device. The method includes: firstly, acquiring visual cue text of a target object to be detected in a target image; wherein the target image is an environmental image of the vehicle's surroundings; then, detecting the target object from the target image based on the visual cue text using an OVD model to obtain a first detection result, and performing re-frame filtering processing on it to obtain a second detection result of the target object; next, filtering out the data with detection errors in the second detection result of the target object to obtain the target image data processing result. Since this application introduces OVD and adopts a multi-filtering strategy (such as an image data processing strategy that combines the CLIP model and the VLM model) to filter out the data with detection errors of the target object in the target image (such as data with shape and text detection errors), it improves the accuracy of the target detection result, the user's riding experience, and the vehicle's driving safety.

This invention discloses an image detection method for ancient and modern Chinese text in complex backgrounds based on text focusing, belonging to the field of computer vision and text detection technology. The method includes: acquiring multi-scale feature maps of the input image through a feature extractor; highlighting text regions using an adaptive text region focusing module through axial pooling and shape self-calibration; fusing features at different scales via a gating mechanism of a dynamic fusion upsampling module; generating probability maps, threshold maps, and approximate binary maps using a differential binarization module to accurately segment the text; and finally extracting text bounding boxes based on the approximate binary maps. This invention effectively improves the detection accuracy and robustness of text in ancient and modern Chinese scenes, significantly enhancing its adaptability to complex backgrounds, diverse fonts, and text with extreme aspect ratios. It can be widely applied in fields such as ancient book digitization, inscription preservation, and intelligent management of cultural heritage.

The application discloses a handwriting text neatness evaluation method, device and equipment and a storage medium. The method comprises the following steps: obtaining an image of a handwriting text to be evaluated; detecting the image by using a pre-trained character detection model to obtain a character detection result; the character detection result comprises the positions of character detection regions in lines; the line spacing of each line of characters is calculated according to the positions of the character detection regions; a standard character handwriting line spacing threshold range is obtained; the number of line spacings in the standard character handwriting line spacing threshold range is counted and recorded as the number of qualified line spacings; a first score of the handwriting text is obtained according to the number of qualified line spacings; and the neatness score of the handwriting text is evaluated according to the first score. The application provides an evaluation method of neatness, can evaluate the overall aesthetic degree of the handwriting text as a whole, and makes the handwriting text evaluation method more comprehensive.

The application provides a machine-generated text detection method, system, device and medium, and belongs to the field of natural language processing. The method comprises the following steps: performing sentence segmentation on an original text; generating a perturbed text through a four-layer structure perception perturbator, wherein the perturbator dynamically selects a perturbation type and intensity through a learnable strategy; reorganizing the text through a pre-trained encoding-decoding model and extracting a feature vector; fusing the multi-source semantic and structural features of the original text, the perturbed text and the reorganized text; and inputting a classifier to output a text detection result in combination with a comprehensive loss function. The problem of low detection accuracy of the text after machine polishing is solved.

This invention provides a text detection method, which includes: acquiring a text to be detected; performing sensitive word detection processing on the text to be detected based on a preset sensitive word tree to obtain a first sensitive word set and a first non-sensitive word set; performing sensitive word detection processing on the text to be detected based on a preset large language model to obtain a second sensitive word set and a second non-sensitive word set; and determining whether the text to be detected contains sensitive words based on the first sensitive word set, the first non-sensitive word set, the second sensitive word set, and the second non-sensitive word set. Due to the characteristics of the large language model, the sensitive word detection processing can take into account the inherent semantic features of the text. Therefore, by integrating the second sensitive word set and the second non-sensitive word set related to the inherent semantic features, as well as the first sensitive word set and the first non-sensitive word set related to the sensitive word tree, the accuracy of text detection can be improved.

This invention discloses a machine-generated text detection method based on frequency domain feature extraction and local frequency filtering. The method involves embedding the input text and constructing an initial frequency domain representation, then building a local frequency filtering module. This module extracts local salient features of the text in different frequency bands through peak detection and frequency band mask generation, forming a multi-band local frequency representation, and constructing a global frequency domain representation that spans both text and models. Text features are fused in both the time and frequency domains, and a unified discriminant representation for machine-generated text detection is constructed through spectrum reconstruction. The local frequency filtering module, spectrum reconstruction, and discriminant network are jointly optimized end-to-end to output prediction results. This invention can effectively identify the potential structural patterns of machine-generated text in the frequency domain dimension, exhibiting high detection accuracy, strong cross-model robustness, and excellent out-of-domain generalization performance. It can be widely applied in scenarios such as content security review, text credibility assessment, and source tracing analysis of content generated by large models.

This invention discloses a method and apparatus for text recognition and drug name extraction on medicine bottle labels based on line scanning imaging. The apparatus includes: a rotating stage for carrying and rotating the medicine bottle; a line scanning camera that acquires line-by-line image data at a constant line frequency while the medicine bottle rotates; an image stitching unit that stitches the line-by-line images in real time according to encoder signals to generate a cylindrical unfolded image, eliminating surface distortion at the hardware level; a text recognition unit that performs OCR text detection and recognition on the unfolded image and outputs a fully preserved text sequence; and a semantic parsing unit that uses a pre-trained large language model to perform semantic-level error correction on the text sequence and extracts the generic name, brand name, and specifications of the drug. This invention actively eliminates distortion and reflection interference through line scanning imaging and achieves semantic-level text understanding by combining a large language model, significantly improving the accuracy and robustness of drug name extraction. It is suitable for automated finished product verification scenarios of cylindrical medicine bottles.

The application relates to the technical field of intelligent decision-making, and discloses a text detection method, a text detection device, equipment and a medium: a to-be-detected text is acquired, multi-modal analysis is performed on the to-be-detected text, and a multi-modal text fingerprint is obtained; when a text verification result of the to-be-detected text represents that all text verifications pass, a text semantic version graph is acquired; text matching is performed on the multi-modal text fingerprint and historical text fingerprints in the text semantic version graph, and a target detection text is obtained; a difference between the to-be-detected text and the target detection text is determined as a semantic difference, and when the semantic difference is key field information and a state of the target detection text is completed, the to-be-detected text is determined as a risk text. The application is applied to a financial scene. Through the multi-modal text fingerprint and the text semantic version graph, subtle information in the to-be-detected text and overall text fingerprint detection are realized, and the accuracy and reliability of risk text detection are improved.

This invention belongs to the field of machine vision, specifically a deep learning-based dot matrix character recognition method, comprising the following steps: 1) acquiring the font and size of characters from an inkjet printer, creating a dot matrix sequence, and generating dot matrix character images in a newly created background image according to the dot matrix sequence; 2) adding random elastic deformation to each character image in the dot matrix character images to obtain a dataset containing multiple character images; 3) fusing each character image as a foreground image with the background image to obtain a fused image; 4) using the fused image to establish an image dataset, constructing a network training model, and training the network model based on the image dataset, inputting the image to be recognized into the network training model for text detection and recognition to obtain the character. This invention establishes a new network structure by inserting a Spatial Variation Network (STN) into a Spatial Convolutional Recurrent Neural Network (CRNN), enabling accurate recognition of dot matrix characters.

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.