Abstract: Video frames that contain text areas are selected from given video frames by removing redundant frames and non-text frames, the text areas in the selected frames are located by removing false strokes, and text lines in the text areas are extracted and binarized.

Abstract: Precise grayscale character segmentation apparatus and method. The precise grayscale character segmentation apparatus comprises an adjustment and segmentation unit for adjusting and segmenting an inputted low resolution text line image undergone coarse segmentation, so as to generate an adjusted character image; a character image binarization unit for generating a binary character image from the character image inputted therein; a noise removal unit for removing noise information in the binary character image generated by the binarization unit; and a final character image segmentation unit for generating a precisely segmented character image from the binary character image from which noise has been removed.

Abstract: A pointing information extraction unit extracts pointing information indicating a pointing position and a pointing time on a slide from a slide file used in a lecture and a video file of a lecture video using a pointing device. A word information generation unit analyzes a text sentence extracted from the slide file to generate a word information file indicating a word and a position thereof. A word pointing information generation unit estimates a word closest to the pointing position on the slide to generate a word pointing information file with the pointing time assigned. A fill-in-the-blank word extraction unit extracts a word having a pointing time equal to or longer than a predetermined time from the word pointing information as a fill-in-the-blank word file. A fill-in-the-blank test question is generated by setting the fill-in-the-blank word of the slide information as a blank region.

Abstract: A document layout analysis program capable of extracting an appropriate set of text blocks from a given document image even in the case where the document layout is so complicated that conventional extraction methods with a single extraction condition would not work well. A plurality of different extraction conditions are stored in an extraction condition memory for use in extracting text blocks from a given document image. In accordance with those extraction conditions, a text block extractor extracts a plurality of sets of text blocks from the document image. A text block consolidator produces a consolidated set of text blocks by performing character recognition on each extracted text block, evaluating validity of each text block based on a result of the character recognition, and selecting most valid text blocks from among the plurality of sets of text blocks.

Abstract: A grayscale character dictionary generation apparatus, comprising a first synthetic grayscale degraded character image generation unit for generating first synthetic grayscale degraded character images using binary character images inputted therein; a clustering unit for dividing each category of the first synthetic grayscale degraded character images generated by the first synthetic grayscale degraded character image generation unit into a plurality of clusters; a template generation unit for generating template for each of the clusters; a transformation matrix generation unit for generating transformation matrix in relation to each of the templates; and a second synthetic grayscale degraded character dictionary generation unit for obtaining character feature of every grayscale degraded character of each of the clusters using the transformation matrix, and for constructing eigenspace of each category of the synthetic grayscale degraded character, which is the second synthetic grayscale character dictionary.

Abstract: A method and apparatus for generating a degraded character image at various levels of degradation automatically is presented in this invention. The method comprises rendering the character image on a scene plane; translating and rotating the scene plane according to various parameters; determining a projection region of the character image on an image plane according to various parameters; generating a pixel region mask; and generating a final degraded image by super-sampling. Thus various degraded character images are generated on various conditions of degradation. The generated synthetic characters can be used for performance evaluation and training data augmentation in optical character recognition (OCR).

Abstract: A method and apparatus for generating a degraded dictionary automatically is presented in this invention. Herein, a degraded pattern generating means generates a plurality of degraded patterns from an original character image, based on a plurality of degradation parameters. A degraded dictionary generating means generates a plurality of degraded dictionaries corresponding to the plurality of degradation parameters, based on the plurality of degradation patterns. Finally, a dictionary matching means selects one of the plurality of dictionaries which matches the degradation level of a test sample set best, as the final degraded dictionary. In this invention, various degraded patterns can be generated by means of simple scaling and blurring process for establishing degraded dictionaries. Therefore, the invention can be implemented simply and easily.

Abstract: According to an aspect of an embodiment, a method of character recognition out of an image having a frame and a plurality of characters in an area, comprises the steps of: dividing the area into a plurality of partial areas having a plurality of partial images, respectively; providing a template image having a reference frame image; calculating differences between the partial images and the reference frame image of the template image, respectively; calculating misalignment of the image from the template image based on the average of the differences of the partial images and the reference frame image; and recognizing the characters out of the image upon correction of the misalignment.

Abstract: A projection set of geodesic lines which are parallel with each other on a curved surface of a paper face is extracted from an image in which a paper face has been imaged by an image-pickup device, using the paper face contents as a clue; and also a projection set of ruling lines which form a ruled surface corresponding to the curved surface of the paper face is extracted from the projection set of geodesic lines. Then, the curved surface of the paper face is estimated from the projection set of the geodesic lines and ruling lines, and distortion of the image is corrected based on this curved surface of the paper face. If this is done, correspondence with various types of diverse distortions becomes possible, and distortion correction can be performed even when only one part of the paper face appears in the image.

Abstract: A rather expanded binary image and a rather blurry binary image are generated from a multiple-valued image. A ruled line candidate area is extracted from the rather expanded binary image, and the extracted ruled line candidate area is verified using the rather blurry binary image.

Abstract: This invention provides a correcting device and a correcting method for perspective transformation of document images.

Abstract: An image distortion correcting apparatus is provided with an image input section to input an image of a flat rectangular paper surface imaged by an imaging section, as an input image, an imaging position estimating section to estimate a relative imaging position of the imaging section with respect to the paper surface from four vertexes of the rectangular paper surface within the input image, a rectangular paper surface estimating section to estimate four vertexes of the rectangular paper surface within a three-dimensional space based on the imaging position, and an image correcting section to correct a perspective transformation distortion in the paper surface within the input image based on the imaging position and the four vertexes within the three-dimensional space, so as to output an output image.

Abstract: The present invention as disclosed hereby is to provide an apparatus and method for allowing images to correspond to each other and a program therefore to automatically perform or make synchronous correspondence of images. According to the apparatus on the condition that there are OHP files and a moving image photographing a lecture or the like using such OHP files, resemblances between the OHP files and OHP image regions of frame images constituting the moving image is calculated such that the correspondence processing between the OHP files and the frame images can be performed based on the calculated resemblances.

Abstract: A set of straight lines that associate a top parallel geodesic projection positioned at an upper end with a bottom parallel geodesic projection positioned at a lower end, among sets of parallel geodesic projections, is extracted as a set of ruled-line candidate projections as a search target of a set of ruled line projections. A deviation of neighborhood, which is a distance between a cross ratio vector of the ruled-line candidate projection and a cross ratio vector of a neighboring line obtained by shifting the ruled-line candidate projection by a predetermined interval, is calculated for each ruled-line candidate projection. A set of straight lines having the smallest sum total of deviations of neighborhood, in the set of straight lines, which do not intersect with each other, among the sets of ruled-line projection candidates is extracted as a set of ruled line projections by continuous dynamic programming.

Abstract: A three-dimensional curved-surface model can be estimated by using both two-dimensional outlines obtained from a piece of image photographed from the top and a restriction that the paper is rectangular. Then, only the three-dimensional distortion in the image can be corrected based on the obtained three-dimensional curved-surface model.

Abstract: Upon receiving, for example, document data including a character string from outside, a character recognition device detects a line from a line-touching character-string image in which at least one character (such as number, alphabet letter, kana character, and Chinese character) touches (or overlaps) a line in the document data, tentatively removes the line, and estimates a character region. The character recognition device extracts a line-touching character image from the line-touching character-string image (original image) based on the estimated character region. The character recognition device creates a line-added reference character image by adding a quasi-line to a reference character image stored in advance.

Abstract: Herein disclosed is a memo image managing apparatus, system and method which are adapted to manage and retrieve a large amount of memo images. The memo image managing apparatus that is connectable to a portable terminal capable of performing a memo image exchange and a memo image display comprises: a managing server adapted to manage memo images and a character information recognizing server adapted to recognize character information of the memo image. The managing server is comprised of a memo image accumulating section adapted to accumulate memo images and a memo image retrieving section adapted to retrieve a predetermined memo image from the accumulated memo images in response to a request. The character information recognizing server is comprised of a character information recognizing section adapted to recognize character information from the memo image.

Abstract: A document information input apparatus detects a position and an attribute of an area of a real document to be input designated by a user with high accuracy. Based on the detected position and attribute, the document information input apparatus recognizes an image of the area as text information by performing recognition processes suitable for the detected attribute such as character recognition, table recognition and a figure process. Then, the document information input apparatus pastes the resulting information to a pertinent position of an electronic document on a display. As a result, it is possible to input information such as a character sequence, a table and a figure from a real document to an electronic document at high speed and with high accuracy.

Abstract: Learning material information containing a fixed attribute specifying learning materials and a variable attribute by which at least levels specifying an object person of learning are defined, is registered in a database of a learning material management system. When the learning request is received by the system from a learning applicant, with reference to the variable attribute of the learning material information registered in the database, a learning material adapted to a level of the learning applicant according to the learning request is selected to be provided for the learning applicant, and on the other hand, the variable attribute of the learning material information registered in the database is updated according to a learning result of the learning material by the learner.

Abstract: In an apparatus for analyzing a layout of a document, a character candidate element generator generates character candidate elements from black pixel linkage components of a document image. A horizontally oriented line rectangle generator sets a plurality of character candidate elements as a line candidate rectangle, among character candidate elements aligned in horizontal line orientation, when each amount of displacement of the set character candidate elements in a vertical orientation with respect to the horizontal line orientation, is smaller than or equal to a threshold value. A horizontally oriented paragraph-box generator sets a plurality of line candidate elements having approximately the same length as each other in the vertical orientation, as a paragraph candidate element.