32results about "Digital ink recognition" patented technology

Systems and methods for providing answers to digitally inked questions within an electronic document. One system includes an electronic processor configured to receive detected interactions between a touchscreen and a digital pen representing one or more digital strokes within a canvas and determine whether the one or more digital strokes match a predetermined signifier. The electronic processor is also configured to, in response to determining that the one or more digital strokes match the predetermined signifier, determine a digital ink instance included in the canvas based on a position of the one or more digital strokes within the canvas, extract text from the digital ink instance, submit the extracted text to an answer service as a query, receive a text-based answer to the query from the answer service, convert the text-based answer to digital ink, and add the digital ink to the canvas.

An electronic device may: receive an input for searching for note contents stored in a memory, the input may include at least one of a piece of text input through a virtual keyboard or a path of an external object dragged on a display; determine, based on identifying a first piece of text from an input, one or more second pieces of text, which appear to be related to the first piece of text, using a language model representing a relationship between words; determine feature information of one or more paths based on identifying the one or more paths from the input; execute a function for searching a database related to note contents using at least one of the one or more second pieces of text and the feature information; and display a screen providing at least one note content on the display in response to at least one note content matched with the input, identified based on the execution of the function.

The application provides a handwritten mathematical expression recognition model training method, a recognition method and a device. LaTeX sequences are automatically extracted and enhanced from academic PDF files, and then a large-scale and high-real handwritten formula image dataset is generated through rendering and material conversion, and then an end-to-end recognition model is trained. The scheme fundamentally solves the core bottleneck of training data scarcity and diversity deficiency caused by high artificial annotation cost; through the synthesis of realistic handwritten images, the generalization ability and robustness of the model in the real scene are effectively improved; finally, the recognition model realizes a significant breakthrough in performance when processing complex and long sequence formulas, and provides a complete and scalable engineering solution for the large-scale application of the technology.

A method for providing recommendations based on handwritten data in a handwritten document is disclosed. The method includes fetching the handwritten data corresponding to strokes made by a user in the handwritten document on a screen of an electronic device. The method also includes recognizing text associated with the fetched handwritten data by employing a handwriting recognition technique. The method further includes identifying keywords in the recognized text and determining recommendations based on the identified keywords using a named entity linking and / or a knowledge graph. The recommendations include handwritten documents stored in a database and / or weblinks. The method also includes ranking the determined recommendations based on a similarity score of each of the determined one or more recommendations with respect to the identified one or more keywords and rendering the ranked recommendations to the user.

According to an embodiment, an electronic device may comprise: display; memory configured to store one or more instructions, and one or more processors operatively connected to the display and the memory. When the one or more instructions are individually or collectively executed by the one or more processors, the one or more instructions cause the electronic device to, identify an editing object added to a first image based on an editing operation being performed on the first image, determine a size of at least one block based on information of the editing object, based on at least one block of the determined size, determine partitions associated with the editing object, extract at least one significant block in which the editing object is at least partially comprised from the at least one block, and store an integrated block generated based on the at least one significant block.

In some embodiments, an electronic device selects and provides for interaction with handwritten content in a content entry region. In some embodiments, an electronic device generates normalized shapes based on handwritten inputs. In some embodiments, an electronic device identifies actionable text within handwritten content and updates the identified text to be actionable. In some embodiments, an electronic device presents tutorials for text manipulation operations. In some embodiments, an electronic device displays visual feedback of text manipulation operations. In some embodiments, an electronic device selects units of content.

The claimed invention relates to management of documents with branding. The system receives at least one document and detects at least one logo. The system retrieves branding information based on the detected logos and generates folders based on the branding information and a preset threshold. The system then modifies a location of documents based on the branding information and the preset threshold.

A method for determining at least one target action from a set of actions able to be executed on an electronic terminal. Such a method includes: obtaining a current digital context of use of the electronic terminal, including data related to a current use of the terminal; analyzing at least one criterion of similarity between the current digital context and a plurality of previously stored reference contexts, on the basis of the data, delivering at least one context of highest degree of similarity with the current digital context from the reference contexts, referred to as close context; and determining the at least one target action by comparing the current digital context and the at least one close context.

The application discloses a stroke sequence recognition method and device and computer equipment, and mainly lies in that the stroke sequence recognition can be carried out in the daily writing process of students, and the writing problems of students can be corrected in time. The method comprises the following steps: in response to the rotation of a support device of a tablet computer by a preset angle, an image sequence in a learning scene is captured by using a shooting device arranged at one end of the support device; the image sequence is subjected to segmentation processing to obtain a hand image sequence and a writing tool image sequence; the hand image sequence and the writing tool image sequence are respectively subjected to registration to obtain a first motion displacement of a hand and a second motion displacement of a writing tool; and the stroke sequence of a current writing character is determined based on the first motion displacement and the second motion displacement.

The application discloses a handwriting character erasing method based on sentence-level connected domain generation and region perception description feature extraction, which is applied to the field of text segmentation and aims at the problem that the existing technology is difficult to capture the fine visual difference between handwriting and printed matter and the segmentation precision is far from reaching the industrial standard. The application firstly introduces a minimum discrimination unit system taking a sentence-level connected domain as a core to realize accurate conversion of physically discrete components into semantically consistent units by using a union-find set and an adaptive threshold aggregation mechanism; then, feature modeling is carried out by using the physical difference between the inherent variability of handwriting and the structural uniformity of printed matter; finally, attribute mapping is carried out by using the extracted high-dimensional feature vector, and handwriting character erasing is realized based on the mapping result. The application can effectively avoid the misdeletion of printed texts under the scene of stroke interlacing and complex background, reduce handwriting residues, improve erasing integrity and keep the stability of the layout structure of the original document.

A writing device assembly and a writing content identification method are provided. The assembly includes a writing apparatus including a pressure sensing unit, an optical identification unit and a processor, and a writing medium having code points printed thereon. The pressure sensing unit generates a pressure sensing signal when sensing that the writing apparatus contacts the writing medium. The optical identification unit captures a code point image of the code points in the writing medium when the pressure sensing unit generates the pressure sensing signal, analyze the code point image to obtain a code point value, and send the code point value to the processor. The processor identifies a code point type of the code point value, analyzes the code point value to determine a writing trace when the code point type is identified to be a coordinate code, and identifies the writing trace to obtain writing content.

The present disclosure relates to systems, non-transitory computer-readable media, and methods for generating validity notifications including validity classifications utilizing a communication validation machine learning model. In particular, in one or more embodiments, the disclosed systems receive a digital image including a digital communication. Further, in some embodiments, the communication validation system utilizes a communication validation machine learning model to analyze the digital image and determine a validity classification. In one or more embodiments, the communication validation system also generates and provides a notification including the validity classification to a client device.

An illustrative system is configured to render, based on depth data representative of a depth map for a scene and within a 3D image depicting the scene, a telestration element at a 3D position within the scene, the 3D image acquired by an imaging device while the imaging device is aimed at an internal space of a patient; and display, within an additional 3D image acquired by the imaging device while the imaging device is aimed at a different internal space of the patient that does not depict the 3D position within the 3D image where the telestration element was rendered, a directional graphical indicator that points in a direction of the 3D position within the 3D image where the telestration element was rendered.

Title: Writing and Image Capture Device The invention relates to a writing and image capture device (1), comprising: - a pen body (2) extending along a central axis (20), and having two lateral finger positioning zones (21) located on either side of a median sagittal plane (MSP) of the pen body (2) passing through the central axis (20), the body (2) also having a ventral region (22) intended to be oriented towards a writing surface; - writing means (3) received in the body, and comprising a writing tip (31) emerging from the body (2), the tip (31) being coaxial with the central axis (20); - a camera (4) inserted in a housing (24), and fixedly mounted relative to the tip (31), the camera (4) comprising a front lens (41) located near the lateral finger positioning zones (21) fingers, towards the tip of the point (31). Figure for the abbreviation: Fig. 1

The application provides a font generation method, device, electronic equipment and system, and relates to the technical field of word processing. The method comprises the following steps: obtaining a standard stroke sequence of a to-be-generated character, and extracting a first character content feature of the standard stroke sequence; obtaining a first single-character image in a handwritten character image, and inputting the first single-character image into a style feature extraction model to obtain a first writing style feature output by the style feature extraction model; extracting a second character content feature of the first single-character image, and removing the second character content feature from the first writing style feature to obtain a second writing style feature without content information; and inputting the first character content feature and the second writing style feature into a decoder to obtain a first target style character corresponding to the to-be-generated character output by the decoder, wherein the decoder is used for fusing a character content feature and a writing style feature. The technical scheme provided by the application can generate personalized fonts and improve the personalization degree of the fonts.

This invention discloses a high-precision handwritten mathematical formula recognition method, system, device, and storage medium. These are corresponding solutions. Addressing the problem that existing recognition methods over-rely on text-level evaluation metrics, leading to low visual fidelity and an inability to effectively handle structured ambiguities and subtle structural errors, this invention establishes a mapping from the symbolic domain to the visual domain. It measures the image-level matching score by calculating the difference between the predicted formula image and the true formula image, and uses this score as a reward signal for network training, directly updating the parameters of the policy network. This effectively overcomes the problem of multiple codes for a single image, significantly improving the model's accuracy and robustness in recognizing the structure of complex mathematical expressions.

A system and method for generating a sequence of ink points to simulate a handwriting style of an identified user, comprising: receiving (S110) a character string input, each character of the character string having a character label; selecting (S120) the identified user; retrieving (S130), for each character label, a subsequence of a fixed number of ink points; selecting (S140), for each retrieved ink point, a first set of ink point features; generating, for each retrieved ink point, a second set of random ink point features; applying (S160), for each character label, the first and second sets of ink point features to a reverse process of a trained diffusion model of a predetermined number of diffusion steps to generate, at each diffusion step, a version of the second set of ink point features corresponding to the respective diffusion step, to depict a handwriting version of the respective character label until all the character labels have been depicted; concatenating (S170) the depicted handwritten versions of the respective character labels.

This invention relates to a composite functional capacitive sensing chip, comprising: a substrate, a signal processing circuit layer, an interconnect layer, an electrode array layer, and a dielectric layer. The signal processing circuit layer is disposed above the substrate and uses an algorithm to extract fingerprint and handwriting features. The interconnect layer is electrically coupled to the signal processing circuit layer for transmitting capacitance signals. The electrode array layer is arranged in a matrix array to detect changes in capacitance caused by contact pressure and movement trajectory of the contact object; the signal processing layer generates digital information of the fingerprint pattern and handwriting trajectory. The dielectric layer is electrically coupled to the electrode array layer and serves to isolate the electrodes, prevent short circuits, and enhance the capacitance effect; it is formed of a high-dielectric-constant dielectric material, a varistor material, or an insulating material.

A handwritten character beautification method is disclosed. Information about a first handwritten character may be obtained. In addition, a fusion style may be determined based on information about a first level, a handwritten character library, and a target font. The fusion style may indicate a style obtained after the target font and a handwritten font corresponding to a character in the handwritten character library are fused. Finally, a first beautified character corresponding to the first handwritten character is generated based on the information about the first handwritten character and the fusion style. In this way, a handwritten character of a user can be beautified based on the first level selected by the user, and a personal style of the handwritten character of the user can be retained in a finally presented beautification effect of the handwritten character. An electronic device is further disclosed.

An information processing apparatus includes: an input unit capable of detecting handwritten input; a display unit capable of displaying a trajectory of the handwritten input; a text recognition unit which recognizes text based on the handwritten input detected by the input unit; a prediction processing unit which predicts text candidates following the text based on the text recognized by the text recognition unit; a handwriting synthesis unit which synthesizes the text candidates predicted by the prediction processing unit into handwritten characters by combining first generation processing to generate handwritten characters based on tokens, each of which is a set of one or more handwriting strokes, collected from the handwritten input by a user, and second generation processing to generate handwritten characters based on a conversion model prepared in advance for converting text characters to handwritten characters; and a display processing unit which displays, on the display unit, the handwritten characters.

This invention provides a writing method and smart pen based on big data-assisted writing. It receives test data from a writing user, acquires the characters to be trained from the test data, retrieves a training database to perform similarity analysis on the characters to be trained, and obtains the predicted training counts from the deviation state to the standard state for each character to be trained. It then acquires the writing user's historical training data, and obtains adjustment data based on the predicted training counts and actual training counts for each character to be trained in the historical training data. Based on the adjustment data, it adjusts the predicted training counts to obtain the planned training counts for the corresponding characters to be trained. Based on the planned training counts and the baseline writing counts, it obtains a customized writing page corresponding to the writing user. Finally, it customizes the combinations of the characters to be trained in the customized writing page according to a reconstructed lexicon, generating a combined writing page corresponding to the writing user.

The subject matter of this specification can be implemented in, among other things, a method to display a graphical user interface (GUI) of a graphics application, receive a selection of a first granular medium tool (GMT), receive a first input associated with the first GMT, the first input comprising one or more first input values characterizing placement of a first amount of a granular medium (GM) on a virtual drawing canvas, access one or more settings of the first GMT, determine, using the one or more first input values and the one or more settings, a first spatial distribution of the first amount of the GM on the virtual drawing canvas, and display, on the GUI, a first image corresponding to the first spatial distribution of the first amount of the GM.

The application discloses a line drawing interruption identification method and device of an electronic whiteboard, the electronic whiteboard and a medium. The method comprises the following steps: acquiring a to-be-processed image, and performing binaryzation processing on the to-be-processed image to obtain a black-and-white image; performing connected domain segmentation and merging processing on a writing track in the black-and-white image to obtain a connected domain merging image; performing edge detection on the connected domain merging image to obtain an edge image; adopting a preset connected domain analysis algorithm to perform identification and marking processing on connected domains in the edge image to obtain target connected domains; for any target connected domain, scanning a boundary point set in the target connected domain to obtain a line drawing interruption point; and when the scanning of all the target connected domains is completed, recording and marking the identified line drawing interruption points to obtain a target result. The line drawing interruption points are comprehensively scanned and identified, and the accuracy of identifying the line drawing interruption is improved.

An Online Software Platform (OSP) is configured to receive an upload trusted document, input text segments of it, identify from the trusted document a set of requirements for the trusted document, confirm that the requirements are met, and therefore cause to be created a trusted document record that can be used to process a dataset that is received via a network later. The dataset has dataset values that can be used to select a subset of stored digital resource rules. The subset includes a main rule and an override rule, and it can be determined from the trusted document record whether the override rule applies or not. If so, a resource having a value of zero is reported. Else, a resource having a non-zero value is produced by applying the main rule to at least one of the dataset values, and is reported.