Abstract: Root cause estimation for a data set corresponding to customer returns of a product may use a probabilistic model to associate customer-entered product return data with probability distributions relating to possible root causes for the returns. A particular application relates to applying a Bayesian network to customer-selected return reason codes and customer-entered return reason comments to estimate a probability distribution for root causes of a plurality of returns and uncertainties relating to the probability distribution estimation. A bag-of-n-grams can be used to enable the Bayesian network to process natural language portions of the customer-entered product return data. The output of the model and other data relating to the root cause estimation can be conveyed to a seller of the returned products via a user interface.

Abstract: A method, system, and program product for implementing an automated query filtering process for spatial data is provided. The method includes selecting a set of common depth levels for geohash structures. Data indicating results of the selection is stored and a specified depth level of the set of common geohash depth levels is selected. The selected geohash depth level is associated with a spatial column for spatial data to determine a set of geohash depth levels required to generate geohash values. A filter table or index associated with the spatial column is generated based on the selected subset of common geohash depth levels and a relationship between the spatial column, the specified geohash depth level and the filter table is stored within a database. Geohash values for the filter table are generated and a query of the database is executed with respect to the specified geohash depth level, the filter entries, and the filter table.

Abstract: Systems, processes and methods for detecting rotated or angled text in an image based on global text geometry estimations are provided. A method includes, at an electronic device with memory and one or more processors, receiving an image including a plurality of pixels (802); determining, based on the image, one or more pixels of the plurality of pixels included in the image that contain text (804); identifying, based on the one or more pixels that contain text, a plurality of components in the image (810); determining a subset of components based on the plurality of components (814); determining, based on the pixels that contain text of the subset of components, one or more candidate text angles (816); determining a global text angle based on the determined one or more candidate text angles (824); and determining a first plurality of bounding boxes based on the global text angle (830).

Abstract: A method is provided for Optical Character Recognition (OCR). A plurality of OCR decoding results each having a plurality of positions is obtained from capturing and decoding a plurality of images of the same one or more OCR characters. A recognized character in each OCR decoding result is compared with the recognized character that occupies an identical position in each of the other OCR decoding results. A number of occurrences that each particular recognized character occupies the identical position in the plurality of OCR decoding results is calculated. An individual confidence score is assigned to each particular recognized character based on the number of occurrences, with a highest individual confidence score assigned to a particular recognized character having the greatest number of occurrences.

Abstract: The present disclosure includes systems and methods for multimedia image analytic including automated binarization, segmentation, and enhancement using bio-inspired based visual morphology schemes. The present disclosure further includes systems and methods for biometric multimedia content authentication using extracted geometric features and one or more of the binarization, segmentation, and enhancement methods.

Abstract: To include reading design data of a plurality of patterns formed on a sample and characteristic information indicating characteristics of each of the patterns from a storage device, the characteristic information being additionally written in the design data, dividing a pattern formed region of the sample on which the patterns are formed, into a plurality of regions where the characteristics are different from each other on a basis of the characteristic information, calculating parameter information according to the characteristics with respect to each of the regions, where the parameter information is provided for generating a reference image from the design data to be used in an inspection of the patterns, and generating the reference image from the design data on a basis of the calculated parameter information.

Abstract: Systems and methods of integrating message content into a target processing device configured to process input data having a predefined data structure. A messaging server is configured to receive a message from a messaging client device executing a messaging application. An orchestrator device is configured to integrate at least a part of the message content into a target data processing device, receive the part of the message content from the messaging server, and transmit a file derived from the part of the message content to a file processing device. The processing device is configured to transform each received file into a description file comprising a set of predefined keys. The orchestrator device is configured to derive an input file having the predefined data structure from the description file and transmit the input file to the target data processing device for processing of the input file by the target processing device.

Abstract: Image data having text associated with a plurality of text-field types is received, the image data including target image data and context image data. The target image data including target text associated with a text-field type. The context image data providing a context for the target image data. A trained neural network that is constrained to a set of characters for the text-field type is applied to the image data. The trained neural network identifies the target text of the text-field type using a vector embedding that is based on learned patterns for recognizing the context provided by the context image data. One or more predicted characters are provided for the target text of the text-field type in response to identifying the target text using the trained neural network.

Abstract: An image processing apparatus executes a first morphology for a first binary image, to generate a second binary image, specifies a vertical line missing region based on the second binary image, executes a second morphology under a condition different from a condition in the first morphology for the second binary image, to generate a third binary image, acquires pixel information about a region corresponding to the vertical line missing region in the third binary image, and corrects a region corresponding to the vertical line missing region in the first binary image using the acquired pixel information, to generate a fourth binary image.

Abstract: An information processing apparatus includes circuitry configured to acquire first form definition information defining a positional relationship between one or more items and a respective value of the one or more items stored in a memory, recognize and extract a specific item set with a specific character string and a specific value of the specific item from data of a form image based on the first form definition information as a recognition result, and display, on a display, the recognition result and an input reception section used for receiving an input of second form definition information.

Abstract: An information processing apparatus includes a processor configured to: acquire a read image and item information, the read image being an image obtained by reading a paper medium including plural documents, the item information being information of plural items specified by a user from among plural items contained in the documents; extract plural character strings from the read image, each character string being associated with the corresponding one of the items included in the item information; in response to extracting the character strings associated with the item information from the read image, set a split position, the split position being a position at which to split out a portion of the read image as a set of documents, the portion being a portion of the read image from a page where the extracting has begun to a page containing the last extracted character string; and output the read image split in accordance with the split position.

Abstract: An information processing apparatus includes a processor. The processor is configured to receive first image data representing a document, and generate, by processing corresponding to appearance characteristics of the document, second image data not representing information of a deletion target out of information represented in the first image data but representing information other than the information of the deletion target.

Abstract: A computing system includes a vision-based user interface platform to, among other things, analyze multi-modal user interactions, semantically correlate stored knowledge with visual features of a scene depicted in a video, determine relationships between different features of the scene, and selectively display virtual elements on the video depiction of the scene. The analysis of user interactions can be used to filter the information retrieval and correlating of the visual features with the stored knowledge.

Abstract: Systems for item validation and image evaluation are provided. In some examples, a system may receive an instrument and associated data. The instrument may be received and a user profile may be retrieved. The user profile may include a plurality of previously processed instruments that have been determined to be valid and/or authentic. The instrument may be compared to the plurality of previously processed instruments to determine whether one or more elements of the instrument being evaluated match one or more corresponding elements of the plurality of previously processed instruments. Matching or non-matching elements may be identified. In some examples, one or more user interfaces may be generated displaying the instruments and including any highlighting or enhancements identifying matching or non-matching elements.

Abstract: An image processing method to generate a layout of searchable content from a physical document. The method includes generating extracted content blocks in the physical document, generating, based on a bounding box of a text block, a layout rectangle that identifies where machine-encoded text is placed in the layout of the searchable content, generating, based on a bounding box of a non-text block, an avoidance region that identifies where the machine-encoded text is prohibited in the layout of the searchable content, generating, based on the layout rectangle and the avoidance region, a draft layout of the searchable content, and iteratively adjusting a point size of the machine-encoded text in the draft layout to generate the layout of the searchable content.

Abstract: In a case where setting of a file name is performed for a scanned image by using OCR processing results of the scanned image, it is made possible to perform OCR processing for text blocks having a strong possibility of being set as a file name.

Abstract: An information processing device includes: a split printing information setting unit setting split printing information, the split printing information including split position information about a split position in an original image when splitting the original image into a plurality of sub-images and printing the sub-images and overlap area information about an overlap area between the sub-images when pasting together the plurality of sub-images that are printed; and an instruction document preview generation unit generating an instruction document preview based on the split printing information, the instruction document preview being a print preview of a work execution instruction document for pasting together the plurality of sub-images that are printed.

Abstract: An image captioning system and method is provided for generating a caption for an image. The image captioning system utilizes a policy network and a value network to generate the caption. The policy network serves as a local guidance and the value network serves as a global and lookahead guidance.

Abstract: A method of computerized presentation of a plurality documents is disclosed. There is at least one original document with at least one original document page, and an addendum document with at least one addendum document page. A first selection of the at least one original document is received. There is a page sequencing array defined by an arrangement of each original document. A second selection of the addendum document is received. Each of the at least one addendum document page is correlated to an original document page. A document set is generated using the first selection and the second selection. For each addendum document in the document set, a priority identifier is determined. A document set view is generated from the document set with the original document pages and the addendum document pages, and is defined by an ordered page selection according to the page sequencing array.

Abstract: Systems and methods detect simple user gestures to enable selection of portions of segmented content, such as text, displayed on a display. Gestures may include finger (such as thumb) flicks or swipes as well as flicks of the handheld device itself. The used finger does not occlude the selected text, allowing users to easily see what the selection is at any time during the content selection process. In addition, the swipe or flick gestures can be performed by a non-dominant finger such as a thumb, allowing users to hold the device and make the selection using only one hand. After making the initial selection of a target portion of the content, to extend the selection, for example to the right, the user simply swipes or flicks the finger over the touchscreen to the right. The user could also flick the entire device in a move gesture with one hand.

Abstract: Described systems and methods allow the automatic extraction of structured information from images of structured text documents such as invoices and receipts. Some embodiments employ optical character recognition (OCR) technology to extract individual text tokens (e.g., words) and token bounding boxes from a document image. A feature vector of each text token comprises a first part determined according to a character content of the text token, and a second part determined according to an image content of the token's bounding box. A neural network classifier produces a label indicative of a type of information (e.g. “billing address”, “due date”, etc.) carried by each text token. In some embodiments, documents are linearized by ordering text tokens in a sequence according to a reading order of a natural language (e.g., English, Arabic) in which the respective document is formulated. Token feature vectors are fed to the classifier in the order indicated by the token sequence.

Abstract: An example of apparatus includes a memory to store a first image of a document and a second image of the document. The first image and the second image are Memory captured under different conditions. The apparatus includes a processor coupled to the memory. The processor is to perform optical character recognition on the first image to generate a first output dataset and to perform optical character recognition on the second image to generate a second output dataset. The processor is further to determine whether consensus for a character is achieved based on a comparison of the first output dataset with the second output dataset, and generate a final output dataset based on the consensus for the character.

Abstract: Unlocking digital content embodied in digital readable form on a digital media carrier includes receiving a scanned image of a page from scanning a physical copy of content, evaluating the scanned image; and if the scanned image corresponds to a selected page of the digital content, unlocking the digital content.

Abstract: Particular embodiments described herein provide for an electronic device that can be configured to capture an image on a display, where the image includes at least one user interface element and is part of an application, create a screen signature of the image, determine an exploration strategy for the image based on the screen signature, and perform the exploration strategy on the image. The image can be abstracted to create the screen signature and the exploration strategy includes interacting with each of the at least one user interface elements.

Abstract: Techniques are disclosed for identifying asides within a document, and detecting a display order of contents based of the identified asides. In a document, an “aside” represents a content region of the document that is distinct from the main content regions, and may be visually distinguishable from the main content region. In an example, a document is received, where the document lacks identification of asides. The document is analyzed to identify asides within the document. A display order of contents within the document is then determined, based on the identified asides. For example, in the display order, the asides are ordered between two segments of the main content and/or at a beginning or an end of the main content, but may not be ordered to be embedded in between a segment of the main content. The document is displayed in accordance with the display order.

Abstract: An image processing device includes: an image classifying section which, through a convolutional neural network, classifies each pixel of input image data as expressing or not expressing a handwritten image to calculate a classification probability of each pixel, the classification probability being a probability that the handwritten image is expressed; a threshold setting section which sets a first threshold when removal processing to remove the handwritten image is performed and a second threshold which is smaller than the first threshold when emphasis processing to emphasize the handwritten image is performed; and an image processor which adjusts a gradation value of pixels with a classification probability no smaller than the first threshold to remove the handwritten image when the removal processing is performed and adjusts the gradation value of pixels with a classification probability no smaller than the second threshold to emphasize the handwritten image when the emphasis processing is performed.

Abstract: Disclosed is a new document processing solution that combines the powers of machine learning and deep learning and leverages the knowledge of a knowledge base. Textual information in an input image of a document can be converted to semantic information utilizing the knowledge base. A semantic image can then be generated utilizing the semantic information and geometries of the textual information. The semantic information can be coded by semantic type determined utilizing the knowledge base and positioned in the semantic image utilizing the geometries of the textual information. A region-based convolutional neural network (R-CNN) can be trained to extract regions from the semantic image utilizing the coded semantic information and the geometries. The regions can be mapped to the textual information for classification/data extraction. With semantic images, the number of samples and time needed to train the R-CNN for document processing can be significantly reduced.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating, based on a portrait image, a foreground image mask to indicate foreground pixels of the portrait image; identifying a percentage of white or near white pixels in the foreground by using the foreground image mask and pixel colors in the portrait image; determining whether the percentage of white or near white pixels in the foreground is larger than a predefined threshold; in response to determining, triggering identification of edge pixels in a background of the portrait image; adjusting white background pixels to add shadows by darkening the white background pixels; and adjusting the white or near white pixels in the foreground by darkening the white or near white pixels.

Abstract: Aspects of the present invention disclose a method, computer program product, and system for image processing. The method includes one or more processors generating a first set of binary images from a first image based on a first color attribute value range associated with a textual object presented in the first image. The method further includes one or more processors recognizing a first set of candidates for the textual object from the first set of binary images. The method further includes one or more processors determining a first appearance frequency of a first candidate in the first set of candidates. In response to determining that the first appearance frequency exceeds a first frequency, threshold he method further includes one or more processors determining that the first candidate is a first recognition result for the textual object in the first image.

Abstract: Systems and techniques are provided for generating a social asset from an electronic publication. The system includes providing a template having a set of reserve spaces for elements. The system receives an electronic publication containing elements including images and text passages. The system assigns images from the publication to each of the reserve spaces for images including assigning a first image from the publication to a first one of the reserve spaces for an image. The system chooses a first one of the text passages for associating with the first image. The system selects a portion of less than all of the first text passage. The system generates a social asset by processing the set of reserve spaces to automatically move forward in an animated manner wherein the selected portion of the first text passage superimposes a portion of the first image.

Abstract: Methods, systems, and computer program products for vision-based cell structure recognition using hierarchical neural networks and cell boundaries to structure clustering are provided herein. A computer-implemented method includes detecting a style of the given table using at least one style classification model; selecting, based at least in part on the detected style, a cell detection model appropriate for the detected style; detecting cells within the given table using the selected cell detection model; and outputting, to at least one user, information pertaining to the detected cells comprising image coordinates of one or more bounding boxes associated with the detected cells.

Abstract: Systems and methods are described for generating a machine learning model for multi-modal feature extraction. The method may include receiving a document in a digital format, where the digital format comprises text information and image information, performing a text extraction function on a first portion of the document to produce a set of text features, performing an image extraction function on a second portion of the document to produce a set of image features, generating a feature tree, wherein a plurality of nodes of the feature tree correspond to the set of text features and the set of image features, and generating an input vector for a machine learning model based on the feature tree. In some cases, the feature tree may be generated synthetically, or modified by a user prior to being converted into the input vector.

Abstract: Methods and systems are provided for managing identifying information for an entity. The identifying information of the entity embedded in or associated with a digital image is detected, wherein the identifying information is selected from the group consisting of: text information and image information corresponding to one or more features of an entity. The text information may be removed from the digital image. The image information may be replaced with one or more computer generated synthetic images, wherein the computer generated synthetic images are based on a natural appearance of the digital image. The synthetic content, which may be generated by a GAN, is based on a natural appearance of the image. The medical image may also contain PHI in text-based fields associated with private tags/fields, which are automatically identified and removed using the systems and methods provided herein.

Abstract: An image inspecting apparatus includes a reader that reads an image on a recording material formed in an image forming apparatus and generates read image data and an image analyzer that performs analysis to determine abnormality for the read image data by using a threshold value and creates an analysis result. The image analyzer makes pixels constituting the read image data a target pixel sequentially and performs determination of abnormality for the target pixel by using the threshold value calculated by using a threshold value calculating method. The threshold value calculating method includes a plurality of threshold value calculating methods, and a first threshold value calculating method is switched to other threshold value calculating method correspondingly to a number of pixels included in a region for calculating a threshold value.

Abstract: A method of extracting information from a flowchart image comprising a plurality of closed-shaped data nodes having text enclosed within, connecting lines connecting the plurality of closed-shaped data nodes and free text adjacent to the connecting lines includes receiving the flowchart image, detecting the closed-shaped data nodes, localizing the text enclosed within the closed-shaped data nodes, and masking the localized text.to generate an annotated image. Lines in the annotated image are the detected to reconstruct them as closed-shaped data nodes and connecting lines. A tree frame with the plurality of closed-shaped data nodes and the connecting lines is extracted. The free text is then localized. Chunks of the free text oriented and positioned proximally together are assembled into text blocks using an orientation-based two-dimensional clustering.

Abstract: A method for video content searching is provided. The method accesses video content and segments the video content into a plurality of frames. The method identifies one or more characteristics for at least a portion of frames of the plurality of frames and determines time frames for each characteristic of the one or more characteristics within the portion of the frames. The method generates frame keywords for the plurality of frames based on the one or more characteristics. The method assigns at least one frame keyword to each time frame within the portion of the frames and generates an index store for the video content. The index store is generated based on the frame keywords and assigning the at least one frame keyword to each frame. The index store includes the frame keywords and time frames assigned to the frame keywords.

Abstract: Disclosed herein are a system and method for generating a floating button widget on a host web site. A popup widget may be generated and appear next to the floating button widget on the host website. The floating button widget is implemented via a code snippet integrated into a source code of the host web site. When the integrated code snippet is executed, an external call to an application programming interface (API) via the Internet is made and subsequently generates the floating button widget and/or popup widget on an interface (i.e., a web page) of the host web site.

Abstract: Provided is a method of encoding an image, the method including: obtaining a plurality of patches from the image; obtaining a plurality of transform coefficient groups respectively corresponding to the plurality of patches; inputting, to a machine learning model, input values corresponding to transform coefficients included in each of the plurality of transform coefficient groups; quantizing transform coefficients corresponding to the image by using a quantization table output from the machine learning model; and generating a bitstream including data generated as a result of the quantizing and information about the quantization table.

Abstract: A computer system for generating a labeling term for a set of data entries may include one or more processors having instructions to obtain a set of data entries and identify a set of unique terms. The program instructions further include instructions to determine a frequency of the unique terms and select a first a subset of unique terms based on the frequency. The program instructions further include instructions to form a set of exclusive groups using the unique terms in the first subset and select a second subset of exclusive groups according to a frequency of each exclusive group. The program instructions further include instructions to form distinct terms from the second subset of exclusive groups and designate a label to a set of data entries using the distinct terms. A computer program product and method corresponding to the above computer system are also disclosed herein.

Abstract: The subject matter of this specification can be implemented in, among other things, a method including identifying one or more blocks in an electronic image that depicts text characters. The method includes identifying one or more text blocks among the blocks that depict the text characters. The method includes identifying a text contrast value for each of the text blocks. The method includes identifying a type for each pixel in each of the text blocks based on the text contrast value. The method includes determining, for each pixel in each of the text blocks, a brightness for the pixel based on the identified type. The method includes storing, in at least one memory, the electronic image including the determined brightness for each pixel in each of the text blocks.

Abstract: Visibility of a description on a description field hidden by a presenter is to be ensured while maintaining a positional relationship between the presenter and the description on the description field. The moving image data obtained by imaging a state where the presenter is giving the description onto the description field is processed to determine the description portion. Display data for displaying each of portions determined to be the description portion as a description is generated and superimposed on moving image data. For example, a difference value for each of pixels between a current frame image and a reference frame image is extracted, a group including a series of consecutive pixels having a difference value being a threshold or more is grasped and then, whether or not the group is a description portion is determined for each of the groups.

Abstract: Images related to one or more attacks to a service provider system may be analyzed to improve the security of the service provider system. Each of the images may be segmented into multiple segments. Each of the segments is analyzed independently to determine whether the segment includes obfuscated data and if so, which one of the data obfuscation techniques was used to generate the obfuscated data. Additional information regarding the obfuscated data may be derived from other segments that include unobfuscated data and from the metadata of the image. A data restoration algorithm may be configured accordingly to restore the obfuscated data. The restored data, as well as a context derived for the image, may be used to adjust one or more security parameters of the service provider system to improve the security of the service provider system.

Abstract: An image processing apparatus includes a controller configured to execute a process on a result of reading on a plurality of documents. The controller obtains document data of a plurality of pages generated by reading the plurality of pages, executes detection of a heading region corresponding to a headline of a document on the obtained document data of the individual pages, and makes a determination as to whether reading order of the documents is appropriate by estimating the anteroposterior relationship among the pages based on presence or absence of the heading region in the document data of the individual pages.

Abstract: A system for managing documents, comprising: interfaces to a user interface, proving an application programming interface, a database of document images, a remote server, configured to communicate a text representation of the document from the optical character recognition engine to the report server, and to receive from the remote server a classification of the document; and logic configured to receive commands from the user interface, and to apply the classifications received from the remote server to the document images through the interface to the database. A corresponding method is also provided.

Abstract: The present disclosure provides technical solutions for improving facial image capturing, recognition, and authentication, including: collecting a face image in response to a facial scan instruction (e.g., for facial recognition) using a camera of a mobile terminal; calculating a measure of image brightness (e.g., a luminance value) of the collected face image; enhancing, when a value of the measure of image brightness of the collected face image is less than a first preset threshold, luminance of light that is emitted from a display of the mobile terminal to a target luminance value, and re-collecting a face image using the camera of the mobile terminal and calculating a corresponding value of the measure of image brightness for the re-collected face image; and performing, when the value of the measure of image brightness of the re-collected face image falls within a preset value range, facial recognition based on the re-collected face image.

Abstract: A system for generating a multi-modal summary of a digital document, comprising a processor adapted for: extracting from the document a plurality of graphical elements; generating a set of textual descriptions, each generated for one of the graphical elements and associated therewith; selecting, from the set of textual descriptions and a set of text fragments extracted from the document, a set of representative elements having a highest score computed by applying thereto a score function, where a set of representative elements' score is indicative of a degree by which the set of representative elements represents the document; for each representative element of the set of representative elements, where the element is a textual description of a graphical element of the plurality of graphical elements, replacing the element with the graphical element associated therewith; and generating, using the set of representative elements, another document comprising a multi-modal summary of the document.

Abstract: A device obtains image data associated with a document. Using a first machine learning model, the device determines, for the document, a first classification of one of a plurality of document types and a first confidence score associated with the first classification, and a second classification of one of the plurality of document types and a second confidence score associated with the second classification based on the image data. The device determines a difference between the first confidence score and the second confidence score, compares the difference and a threshold value, and accept the first classification of the document when the difference satisfies the threshold value.

Abstract: Identifying insect species integrates image processing, feature selection, unsupervised clustering, and a support vector machine (SVM) learning algorithm for classification. Results with a total of 101 mosquito specimens spread across nine different vector carrying species demonstrate high accuracy in species identification. When implemented as a smart-phone application, the latency and energy consumption were minimal. The currently manual process of species identification and recording can be sped up, while also minimizing the ensuing cognitive workload of personnel. Citizens at large can use the system in their own homes for self-awareness and share insect identification data with public health agencies.

Abstract: In implementations of grid layout determination from a document image, a computing device receives a document image of a document that includes document content. The computing device implements a grid layout system that can determine feature elements of the document content in the document, and generate a node tree of bounded elements that represent relationships of the feature elements in the document, where each of the bounded elements is considered in the determination of the grid layout. The grid layout system can generate a containment model that includes the node tree of the bounded elements. The grid layout system can then determine a column layout of the columns in the document based on the containment model, which includes calculating a quantity of the columns, and also determine a row layout of the rows in the document based on the containment model, which includes calculating a quantity of the rows.

Abstract: An image processing system adapted to binarize images is provided. The system includes a component detector configured to receive an image and detect a plurality of components in the image. The components are detected based on a content of the image. Further, the system includes a logical splitter configured to split the image into a plurality of windows based on the plurality of components. The plurality of windows is of varying window sizes. In addition, the system includes a threshold detector configured to determine a binarization threshold value for each window. The system also includes a binarization module configured to binarize a plurality of component images based on the corresponding binarization threshold values of the component. Furthermore, the system includes a logical integrator configured to generate a binarized image. The binarized image is a logically integrated image comprising the plurality of component images.