Abstract: This specification discloses a system of using a neural network to distinguish text and pictures in an image and the method thereof. Using the knowledge of text recognition learned by the neural network in advance, images data of color brightness and gray levels in an image block are processed to generate a greatest text faith value. The system determines the text status of the image block by comparing a text threshold with the greatest text faith value. If the greatest text faith value is larger than the text threshold, then the image block is determined to contain text pixels; otherwise, the image block contains purely picture pixels. This achieves the goal of separating text and pictures in an image.

Abstract: A method for recognition of a handwritten pattern comprises the steps of forming (4) a representation of the handwritten pattern, forming (6) at least two subconfigurations by dividing the representation of the handwritten pattern, and processing the subconfigurations. The step of processing comprises the steps of comparing (8) each subconfiguration with reference configurations, selecting (10) at least one subconfiguration candidate for each subconfiguration among the reference configurations based on said step of comparing, and determining (12) at least one candidate pattern consisting of one selected subconfiguration candidate for each subconfiguration. The method further comprises the steps of comparing (14) the representation of the handwritten pattern to the candidate pattern, and computing (16) a cost function in order to find a closest matching candidate pattern.

Abstract: A system and process for determining the location of a captured image from a larger image is described. Using a list of determined locations, the system is able to determine the best or most likely path of a pen tip.

Abstract: An object extraction method comprising the steps of comparing each portion of first image data obtained by capturing an extraction object with another portion of the first image data or second image data, thereby generating shape data as an image representing a provisional region of the extraction object in the first image data, and matching a contour of the shape data to a contour of the extraction object by using the first image data and the provisionally generated shape data.

Abstract: Methods of organizing a series of sibling data entities in a digital computer are provided for preserving sibling ranking information associated with the sibling data entities and for attaching the sibling ranking information to a joint parent of the sibling data entities to facilitate on-demand generation of ranked parent candidates. A rollup function of the present invention builds a rollup matrix (126) that embodies information about the sibling entities and the sibling ranking information and provides a method for reading out the ranked parent candidates from the rollup matrix in order of their parent confidences (141). Parent confidences are based on the sibling ranking information, either alone or in combination with n-gram dictionary ranking or other ranking information.

Abstract: Disclosed are embodiments of systems and methods for eliminating or reducing the distortion in a scanned image. In embodiments, the image is segmented into foreground and background pixels. Foreground pixels may be grouped into “letters.” Using index-based searching, “letters” may be grouped into “words” and “words” may be grouped into baselines. One or more dominant baselines may be selected and the characteristics of the dominant baseline or baselines may be used to unwarp the image.

Abstract: An electronic ink object is preprocessed for improving handwriting recognition results. An electronic ink object, possibly having multiple lines of non-horizontal ink, is preprocessed by segmenting or separating the electronic ink object into one or more lines, computing a writing angle of each line, and rotating each line so that it is horizontal.

Abstract: When encoding each element in numeral string data, a code-string data creating unit creates code string data by assigning a code to a focused element to be encoded based on a magnitude relation between the focused element and an element adjacent to the focused element. An control unit stores primary data and the code string data as registration information in an associated manner into a storage unit. When searching for the primary data, a data extracting unit and the code-string data creating unit create code string data from image data, and a searching unit searches data stored in the storage unit based on the created code string data as search information for data having registration information that matches the search information.

Abstract: A character recognition method that is robust under an unknown illumination condition is provided. An apparatus for realizing such robust character recognition includes plural different binarization, means for synthesizing character sub-image candidates that have been obtained from the binarization units, and means for analyzing character sub-image candidates and for recognizing an image as a character string consisting of character sub-image candidates.

Abstract: The present invention describes a process for enhancing optical recognition of text in scanned documents. Prior to performing optical recognition for identification of text in scanned documents, a preprocessing algorithm identifies locations of noncontiguity in character strokes. The gaps created by noncontiguous character strokes are selectively filled with non-white or black pixels for enhanced character recognition. The process may assess noncontiguity on a bit-by-bit basis or, to reduce the number of operations, on a byte-by-byte basis.

Abstract: An information distribution system configured to deliver various types of content provided by an information distributor to information receivers through a network and transmitting the content to be distributed converted to colors, color values, or color digital values. By converting the content to colors, color values, or color digital values, it is possible to reduce the amount of information transmitted. Due to this, it becomes possible to shorten the time required for distribution of content and to improve practicality. Further, it becomes possible to reduce the distribution costs.

Abstract: An apparatus includes a content acquisition unit configured to acquire content data contained in image data, an extraction unit configured to extract a keyword from the image data, a setting unit configured to set acceptance or rejection of modification of the keyword according to a keyword extracted by the extraction unit, and a storage unit configured to store the data of the content, the keyword, and the setting of acceptance or rejection of modification in association with each other.

Abstract: Techniques for converting analog medical video data to digital objects include receiving a digital video signal. The signal is produced by converting an analog video signal from playing a legacy analog medical video medium on an appropriate analog video player. Without human intervention, a first portion of a video frame of the digital video signal is determined where characters are imaged onto the analog medical video by the legacy system. The first portion of the video frame is processed in a video optical character recognition process to generate first character data. Non-video descriptive data associated with the analog medical video data is determined based on the first character data. Digital video data based on the digital video signal is stored in association with the non-video descriptive data. These techniques allow one or more extensive analog medical video libraries to be converted quickly and at low cost in human labor.

Abstract: In an image processor for a character image included in an image data to be processed, character codes are recognized, and character sizes of the characters in the character image are recognized. A font data is selected in a plurality of font data of different sizes stored in a storage device, so as to match with the recognized character codes, the recognized font sizes and a magnification of the image data. The selected font data are outputted. Thus, the character image is reproduced by using the recognized character codes. Alternatively, in a saving mode where image data of N pages are outputted in M sheets of recording medium, wherein N is not equal to M, character codes are recognized in a character image included in N pages of image data to be processed. Then, a font data is selected in a plurality of font data, so as to match with the recognized character codes.

Abstract: A process and system are provided for determining a most likely combination of characters. In particular, character data which includes information indicative of at least one handwritten character is obtained. The character data includes at least one set of segmentation points for the handwritten character. Then, a score can be provided for each particular character of a set of previously stored characters based on a comparison between the character data and the previously stored particular character. In addition, it is possible to compare visual aspects of the handwritten character to visual aspects of each of the previously stored characters for determining likely characters. Also, a Fisher Matching procedure can be used on the character data to ascertain the likely characters. A plurality of handwritten characters can include a first character that is connected to a second character of the handwritten characters via a ligature.

Abstract: A labeling process unit groups a continuous black pixel area as one group in the binary image data read by an image input device, and extracts the group bounding rectangle information about the group. A row extracting process unit extracts row rectangle information from the position information about the extracted group bounding rectangle. An overlap integrating process unit determines the overlap between the group bounding rectangles contained in the extracted row rectangle, and performs an overlap integrating process of integrating overlapping groups into one group. The ratio of the number of group bounding rectangles contained in the row rectangle before performing the overlap integrating process to the number of the group bounding rectangles contained in the row rectangle after performing the overlap integrating process is obtained, and the language of the characters written in the original is determined based on the difference in ratio.

Abstract: A method for recognition of a handwritten pattern comprises the steps of forming (4) a representation of the handwritten pattern, forming (6) at least two sub-configurations by dividing the representation of the handwritten pattern, and processing the subconfigurations. The step of processing comprises the steps of comparing (8) each subconfiguration with reference configurations, selecting (10) at least one subconfiguration candidate for each subconfiguration among the reference configurations based on said step of comparing, and determining (12) at least one candidate pattern consisting of one selected subconfiguration candidate for each subconfiguration. The method further comprises the steps of comparing (14) the representation of the handwritten pattern to the candidate pattern, and computing (16) a cost function in order to find a closest matching candidate pattern.

Abstract: A method, medium, and system extracting text, including filtering a text domain image using a stroke filter, determining a color polarity of the text by using a response value of the stroke filter, binarizing the response value of the stroke filter, and expanding a local domain by using a binary domain generated by the binarization.

Abstract: A raster image path architecture having the capacity for supporting the rendering and output of a device-independent grayscale raster image, while also offering the capacity for supporting the rendering and output of a device-dependent binary raster image, thus offering the advantages of outputting a device-independent grayscale raster image while preserving the performance and image quality advantages of a conventional binary raster image path architecture.

Abstract: A mobile device, system, method, and software for communicating with the internet utilizing a written universal resource locator (URL). A camera unit is used to receive a raw visual light image containing a written URL, the raw visual light image is converted to an electronic image, and the device locates glyphs of at least one particular standardized set of URL characters in the electronic image, for example glyphs corresponding to www. Then the URL characters are extracted from the electronic image, the URL is sent in a request signal to a web server, and in response an internet site is presented. The mobile device includes initiation means for sending an instruction to obtain a raw visual light image that includes glyphs of at least one particular set of characters, such as www, and further includes a camera, a display, and an internet interface. The mobile device processes an electronic image signal provided by the camera, in order to obtain the web site signal from the internet interface.

Abstract: An electronic device (1100) and a method for character segmentation (100) includes an image analyzer (1110) that generates individual character images. The image analyzer binarizes (115) a gray scale image (200) of a horizontal row of characters by using a general threshold method to generate a first image (300). The image analyzer also binarizes (120) the gray scale image using an edge detection method to generate a second image (405). The image analyzer determines (125) a character row region (425) of the second image by using horizontal projection analysis. The image analyzer isolates (130) the character row region of the first image using the character row region of the second image. The image analyzer uses the character row region to generate (135) individual character images. The electronic device may include an image capture device (1105) and a character recognition program (1115).

Abstract: A character recognition processing system includes a character recognition confidence evaluating unit that evaluates whether confidence of character recognition of a plurality of areas are low or high, a character area classification unit that classifies a first area evaluated low by the character recognition confidence evaluating unit into a plurality of components, a character separation unit that separates the components classified by the character area classification unit into a character component and non-character components, according to information relating to a second area evaluated high by the character recognition confidence evaluating unit, and a first character recognition unit that performs character recognition processing for the character component separated by the character separation unit.

Abstract: An automated symbolic recognition system and method includes pre-stored symbolic models to support a three phase symbolic reshaping scheme during the symbol recognition process that includes: (a) deriving dissimilarity level from alphanumeric ID's net topological variation and the integration of each of its arcpoly structural variation(s) signifying a reasonably accurate confidence level for the goodness of recognition, thus establishing a mechanism that derives dissimilarity level (or cost value) between image and database features including shape, size and topological relationship, (b) determining the reshaping or transformation of an arcpoly to another arcpoly by introducing variations to the original arcpoly to alter its shape and orientation and deriving at each step, the new cost value as a function of variation(s) present and imposed, and (c) determining the equivalent representation of an arcpoly by a succession of smaller and adjoining arcpoly(s) in order, or vice versa.

Abstract: An image processing apparatus includes: an extraction unit that extracts a first image and a second image similar to the first image, in a first resolution; and a generation unit that generates an image in a second resolution based on the respective images extracted by the extraction unit and phases of the respective images calculated with precision higher than one pixel in the first resolution.

Abstract: The invention relates to a method for carrying out the imaging processing of data sets of three-dimensional characters obtained from medical examination methods. According to the invention, a separation of the data of relevant anatomic structures as well as of other structures is carried out based on data provided in the form of volume elements. In addition, a spatial representation of at least one structure of interest ensues, and the individual objects of the structure of interest are separated from one another in such a manner that enables them to be repositioned in the spatial representation.

Abstract: A method and computer program product are disclosed for refining character segmentation in an optical character recognition system receiving as input a plurality of candidate objects. Each candidate object below a threshold character width is merged with another candidate object at one or more merge lines to form a composite object. The plurality of candidate objects are preclassified to identify a plurality of composite objects and a plurality of character portions. Proposed split lines are determined for each of the composite objects. Regions are defined within each of the composite objects from the position of the merge and split lines. The defined regions are classified to obtain an associated score for each region. Complete region sets are defined for each composite object, each with an associated set ranking determined from the associated score of the regions comprising the set. The set having the highest ranking is selected.

Abstract: A method and apparatus for reformatting electronic documents is disclosed. In one embodiment, the method comprises performing layout analysis on an electronic version of a document to locate text zones, assigning attributes for scale and importance to text zones in the electronic version of the document, and reformatting text in the electronic version of the document based on the attributes to create an image.

Abstract: An apparatus for embedding a digital watermark in a document image detects circumscribing outer shapes of characters in the document image and sets a plurality of reference lines that extend in the column direction and are spaced apart in the row direction by a basic pitch. The outer shapes include a first outer shape, a second outer shape that neighbors the first outer shape, and a third outer shape that neighbors the second outer shape, and the reference lines include a first reference line located between the first outer shape and the second outer shape, and a second reference line located between the second outer shape and the third outer shape. Control is performed for at least one of the second and third outer shapes so that a distance between the first reference line and an edge of the second outer shape is different from a distance between the second reference line and an edge of the third outer shape, in accordance with digital watermark information to be embedded.

Abstract: Distance d1 between the right edges of A1 and B2, and distance d2 between the right edges of A3 and B4 are calculated. If data to be embedded is 1, one or a combination of a process for increasing the size of B2 in the column direction or decreasing the size of B4 in the column and a process for moving the position of B2 toward B3 or moving the position of B4 toward B3 is executed to meet d1>d2. If data to be embedded is 0, one or a combination of a process for decreasing the size of B2 in the column direction or increasing the size of B4 in the column direction, and a process for moving the position of B2 toward B1 or moving the position of B4 toward B5 is executed to meet d1<d2.

Abstract: Image data is classified to identify the type of the image data using a feature amount of the image data calculated based on the layout (rough spatial arrangement and distribution of texts and photographs or pictures). Based on the result, a region extraction method that is associated with the type of the image data is selected for layout analysis. According to the region extraction method, the image data is divided into regions.

Abstract: An automated symbolic recognition system and method includes multiple representations of each alphanumeric in the database by identifying a finite number of exemplars per alphanumeric symbol.

Abstract: An automated symbolic recognition system and method includes an evidence-based technique to significantly reduce the alphanumeric candidate symbol list.

Abstract: In order to render information contained in images more detectable in the presence of noise, the present invention discloses a system and method to binarize the images in the presence of noise. The method of this invention comprises the steps of applying a dynamic range reducing filter to the digitized image values, obtaining a range reduced image; applying an edge detecting filter to the range reduced image, obtaining a filtered image; and then, adaptively binarizing the digitized image utilizing the corresponding filtered image to obtain an adaptive threshold.

Abstract: An automated symbolic recognition system and method includes the basis for the system's feature set by identifying (a) a set of logical symbols comprising a finite class of arcpolys (lines and arcs and a point) that to the exclusion of the point, each member class has a unique (distinct) orientation, and (b) a set of subclass symbols per logical class of symbol representing a finite subclass of arcpolys (lines and arcs and a point) that to the exclusion of the point, each subclass member has a unique (distinct) extreme points size and/or depth size.

Abstract: The present invention relates to an optical character recognition device (OCR) for reading a form provided with character frames in reading fields, into which a user fills each character. Characteristic vectors are extracted from the character images of each frame. A number of characters decision unit 16, into which the characteristic vectors are input, decides the number of characters filled in one of the character frames. A character separation unit 18 separates each of characters from the character image based on the number of characters decided by the decision unit 16. The character recognition unit 20 then recognizes each of the character. The OCR according o the present invention is able to read the form correctly, in which a plurality of characters are filled in one of the frames.

Abstract: An automated symbolic recognition system and method includes a hierarchical hypothesis-and-verification technique during various stages of the handwriting recognition process, whereby a series of initial assessments are made based on the information availed upon them, and later during processing, they are validated or rejected depending on the degree in which preset milestones were satisfied and are followed by a sequence of alternative hypotheses in the event of failure of the latest hypothesis until they are satisfied.

Abstract: A character overlapping with a line of an entry box is extracted utilizing a dot box. At document input step S2, a document to be recognized is read and converted into electronic data to produce character box data. At electronic data storing step S3, the character box data read at the photoelectric transducing S2 is stored. At character detecting step S5, the character box data is contracted, small points are removed, and the remaining pattern is reversely expanded to produce character data. At entry box detecting step S4, entry box position data is generated from the character box data and character data. At relating step S6, a character pattern entered in each entry box is detected from the character data and the entry box position data. At character recognizing step S7, a character is recognized for the extracted character pattern and the results are stored at step S8.

Abstract: Methods for communicating between an application and an ink divider object (which stores ink strokes to be divided into groups) may include: (a) issuing a divide request to the ink divider object, optionally by the application; (b) in response to the divide request, calling a divide method, which groups the stored ink strokes into one or more groupings of strokes having a first predetermined granularity (e.g., words, lines, paragraphs, sentences, drawings, etc.); and (c) making information regarding the one or more groupings of strokes available to the application. This “information” made available to the application may include, for example, the actual groupings of the strokes, the number of stroke groupings having the first predetermined granularity, machine generated text corresponding to the stroke groupings, or the like. The results of the divide method may be stored in an ink division result object.

Abstract: An automated symbolic recognition system and method includes a three phase symbolic reshaping scheme during the symbol recognition process that includes: (a) deriving dissimilarity level from alphanumeric ID's net topological variation and the integration of each of its arcpoly structural variation(s) signifying a reasonably accurate confidence level for the goodness of recognition, thus establishing a mechanism that derives dissimilarity level (or cost value) between image and database features including shape, size and topological relationship, (b) determining the reshaping or transformation of an arcpoly to another arcpoly by introducing variations to the original arcpoly to alter its shape and orientation and deriving at each step, the new cost value as a function of variation(s) present and imposed, and (c) determining the equivalent representation of an arcpoly by a succession of smaller and adjoining arcpoly(s) in order, or vice versa.

Abstract: An image processing apparatus including a receiving unit, an extraction means, and a conversion means is disclosed. The receiving unit receives a first document data and a second document data obtained by reading respectively a first document and a second document each having at least one drawing. The extraction means extracts a first caption group for identifying the drawing contained in the first document and a second caption group for identifying the drawing contained in the second document. The conversion means converts the first caption group and the second caption group into a third caption group so that a character string in the first caption group does not overlap a character string in the second caption group.

Abstract: The invention relates to a mobile device with a built-in image capture device, and a character recognition function to present the information gathered with the character recognition result. With the mobile device, the character line extraction process is displayed whenever necessary, and the resolution of an image to be inputted for recognition processing is enhanced. Accordingly, it is possible for the operator to select the target character line with ease. In addition, the mobile device has a character recognition ratio improved by the enhancement in resolution.

Abstract: A fast semi-automatic prostate contouring method is provided using model-based initialization and an efficient Discrete Dynamic Contour (DDC) for boundary refinement. The user initiates the process of the preferred embodiment by identifying four (4) points on the prostate boundary, thereby scaling and shaping a prostate model, and then the final prostate contour is refined with a DDC. The method of the present invention has particular application during the pre-implant planning phase of a brachytherapy procedure. However, this method also has uses in any phase of dose planning in the brachytherapy procedure or any other therapy approach.

Abstract: The present invention relates to a retrieval method for searching a first character element string obtained by subjecting a character string to character recognition for a second character element string. The first character element string includes a first character element and the second character element string includes a second character element. A distance relevant to a similarity between the first character element and the second character element is predetermined between the first character element and the second character element. The retrieval method comprises the steps of comparing the distance with a predetermined reference distance, and determining whether the second character element matches the first character element based on a result of the comparison of the distance with the predetermined reference distance.

Abstract: A method for segmenting a small feature in a multidimensional digital array of intensity values in a data processor computes an edge metric along each ray of a plurality of multidimensional rays originating at a local intensity extreme (local maximum or minimum). A multidimensional point corresponding to a maximum edge metric on each said ray is identified as a ray edge point. Every point on each ray from the local extreme to the ray edge point is labeled as part of the small object. Further points on the feature are grown by labeling an unlabeled point if the unlabeled point is adjacent to a labeled point, and the unlabeled point has a more extreme intensity than the labeled point, and the unlabeled point is closer than the labeled point to the local extreme. The resulting segmentation is quick, and identifies boundaries of small features analogous to boundaries identified by human analysts, and does not require statistical parameterizations or thresholds manually determined by a user.

Abstract: Systems and methods for processing text-based electronic documents are provided. Briefly described, one embodiment of a method for processing a text-based electronic document comprises the steps of: comparing at least one word in a text-based electronic document to a native language dictionary to determine whether the at least one word conforms to a predefined rule; for each of the at least one word that does not conform to the predefined rule, fragmenting the at least one word into word fragments; combining at least two consecutive word fragments; and comparing the combination of the word fragments to the native language dictionary.

Abstract: A title is determined from a document image of an unspecified format based upon likelihood of containing a title in each of minimal circumscribing rectangles. The likelihood is determined based upon information obtained during character recognition of the image contained in the minimal circumscribing rectangles. The information includes a font type, a recognition assurance level, character code and so on.

Abstract: A pattern segmentation apparatus and a pattern recognition apparatus can improve the segmentation precision of a character touching pattern. The pattern segmentation apparatus includes a feature amount extraction unit for extracting the feature amount of an image, a feature amount setting unit for setting the feature amount of a category, a feature amount comparison unit for comparing the feature amount of the category with the feature amount of the image, and a segmentation unit for segmenting a portion corresponding to the feature amount of the category from the image based on the comparison result.

Abstract: A method of post-processing character data from an optical character recognition (OCR) engine and apparatus to perform the method. This exemplary method includes segmenting the character data into a set of initial words. The set of initial words is word level processed to determine at least one candidate word corresponding to each initial word. The set of initial words is segmented into a set of sentences. Each sentence in the set of sentences includes a plurality of initial words and candidate words corresponding to the initial words. A sentence is selected from the set of sentences. The selected sentence is word disambiguity processed to determine a plurality of final words. A final word is selected from the at least one candidate word corresponding to a matching initial word. The plurality of final words is then assembled as post-processed OCR data.

Abstract: A system can include a system processor that is programmed or adapted to classify pixels in a document image, and in some instances, to enhance the document image based upon such classification. Image data including pixels is received. For each pixel in some subset of the image data, a feature vector is generated. Each pixel in the subset is classified as text or image based upon its associated feature vector. The classifications can be outputted for storage and/or further processing. Further processing can include pixel level enhancement processing. Such enhancement processing can include application of a sharpening filter to pixels classified as text and of a smoothing filter to pixels classified as image. In some instances, background suppression can be performed upon a subset of the image data prior to classification.

Abstract: Upon synthesizing objects, information bits indicating the types of objects are lost. To solve this problem, this invention provides an image processing apparatus having discrimination means for discriminating a type of object to be rendered, determination means for determining the presence/absence of synthesis of the discriminated object, synthesis means for synthesizing an object and information of the type of object in accordance with the determination result, and processing means for appending information indicating the type of synthesized object to a rendering result obtained by rendering the object to be rendered in units of pixels.