Xerox Corporation

Google Inc.

Hewlett-packard Development Company, L.P.

Other

Dan S Bloomberg

M Margaret Withgott

Lakshmi Dasari

Francine R Chen

Luc Vincent

David L Hecht

Henry W Sang, Jr

Lynn D Wilcox

Ronald M Kaplan

Todd A Cass

Daniel P Huttenlocher

Gary E Kopec

Ashok Chhabedia Popat

Nemanja Spasojevic

Robert Eugene Weltman

Guillaume Raphael Poncin

Per-Kristian G Halvorsen

Robert F Tow

Hui Chao

Dar-Shyang Lee

Daniel Davies

Ramana B Rao

John W Tukey

Steven C Bagley

Leslie T Niles

Daniel C Kowalski

L Prasadam Flores

Ralph Furmaniak

Ray Smith

Thomas P Minka

Philip Andrew Chou

Daniel H Greene

Richard G Stearns

Thomas P Moran

Jan O Pedersen

Daniel Illowsky

Ralph C Merkle

D Austin Henderson, Jr

Z Erol Smith, Iii

John Seely Brown

Robert P Kowalski

Francois-Marie Lefevere

Douglass R Cutting

Marin Saric

Frank Zdybel, Jr

Ramana R Rao

William T Crocca

A signature for a page of text is generated. The signature serves as an identifier of the text page. Positions of words in a text page are determined. Positions of multiple second words in the text page are determined relative to the position of a first word in the text page. A signature value is generated that describes the second word positions relative to the first word position. The signature value is stored. Additional signatures for the text page can be generated, each signature describing positions of other words in the text page relative to a word in the text page for which the signature is being generated. The signatures can be used to compare the text page to another text page and generate a measure of similarity that describes the result of the comparison.

Comparing text pages using image features based on word positions

The present invention relates to systems and methods for analyzing media material having articles continuing across multiple pages. A media material analyzer includes a segmenter and an article composer. The segmenter identifies block segments associated with columnar body test in the media material. The article composer determines which of the identified block segments belong to a continuing article extending across multiple pages in the media material based on language statistics information and continuation transition information.

Media material analysis of continuing article portions

The present invention relates to systems and methods for analyzing media material having a layout. A media material analyzer includes a segmenter and an article composer. The segmenter identifies block segments associated with columnar body text in the media material. The article composer determines which of the identified block segments belong to one or more articles in the media material. The article composer can determine whether candidate block segments belong to a same article based on language statistics information, layout transition information, or both language statistics information and layout transition information. A system for searching media material having a layout over a network is also provided.

Methods and systems for analyzing data in media material having layout

A method and system for storing and generating anti-aliased text and lineart data from compressed document image files, using a MRC model that represents the image as an ordered set of mask/image pairs at resolutions appropriate to the content of each layer. The method and system provide the ability to generate for anti-aliased text data to improve appearance at both high and low resolution, and to avoid baseline jitter of compressed tokens.

Methods for generating anti-aliased text and line graphics in compressed document images

Systems and methods for reorganizing raw image data captured by a camera for improved image processing are disclosed. The method generally includes separately compressing each color component, e.g., RGB, of raw grayscale image data to generate a reorganized grayscale data output file by a first processor, and performing color decoding to generate a color image output by a second processor. The raw grayscale image data may be that of an image of a target captured by a digital camera in an image capturing system. The second processor is separate from the first processor that records the raw grayscale image data of the image captured and may be physically external to the camera. The raw grayscale image data may be a 2N&#xd7;2N array and the reorganized grayscale data output includes 4 N&#xd7;N color-specific arrays where N is a factor of 8.

Reorganization of raw image data for processing

A method and system for storing and generating anti-aliased text and lineart data from compressed document images files, using a MRC model that represents the image as an ordered set of mask/image pairs at resolutions appropriate to the content of each layer. The method and system provide the ability to generate for anti-aliased text data to improve appearance at both high and low resolution, and to avoid baseline jitter of compressed tokens.

A method of aligning a first page image and a second page image is disclosed. The first page image and the second page image are deskewed. Then, the first page image and the second page image are vertically aligned. In particular, a first vertical data set comprising a plurality of first values each first value based on a horizontal scanline of the first page image is generated. Moreover, a second vertical data set comprising a plurality of second values each second value based on a horizontal scanline of the second page image is generated. One of the first and second vertical data sets is dilated. Then, the first and second vertical data sets are cross-correlated to generate cross-correlation data. A maximum value of the cross-correlation data is determined, whereas the maximum value indicates vertical alignment between the first and second page images. Finally, the first and second page images are horizontally aligned.

Method of aligning page images

A system and method for editing and transcribing using a structured freeform editor is provided. The method implemented in the system includes interpreting structure of freeform graphic elements and selectively editing the structure and/or selectively transcribing scribble elements to an editable format.

Method and system for transcribing and editing using a structured freeform editor

This invention provides self-clocking glyph shape codes for encoding digital data in the shapes of glyphs that are suitable for printing on hardcopy recording media. Advantageously, the glyphs are selected so that they tend not to degrade into each other when they are degraded and/or distorted as a result, for example, of being photocopied, transmitted via facsimile, and/or scanned-in to an electronic document processing system. Moreover, for at least some applications, the glyphs desirably are composed of printed pixel patterns containing nearly the same number of ON pixels and nearly the same number of OFF pixels, such that the code that is rendered by printing such glyphs on substantially uniformly spaced centers appears to have a generally uniform texture. In the case of codes printed at higher spatial densities, this texture is likely to be perceived as a generally uniform gray tone.Binary image processing and convolution filtering techniques for decoding such codes also are disclosed, but this application focuses on the codes.

Self-clocking glyph shape codes

A method of document segmentation. Specifically, one embodiment of the present invention discloses a method of document segmentation that performs a plurality of projection profiles of pixel intensities on a document containing a plurality of text lines over a range of angles. A plurality of slope values for a plurality of discrete distances perpendicular to said range of angles is calculated for the plurality of projection profiles. A set of maximum absolute slope values is sorted out from the plurality of slope values. Text lines of first and second type are identified by setting a threshold slope value. Absolute slope values greater than the threshold slope value indicate the plurality of text lines of said first type. Absolute slope values less than the threshold slope value indicate the plurality of text lines of a second type.

Method and system for document segmentation

In a text recognition system that uses a stochastic finite state network to model a document image layout, the computational efficiency of text line decoding is improved. In a typical implementation, the dynamic programming operation that accomplishes decoding uses actual scores computed between two-dimensional (2D) bitmapped character template images and the (2D) bitmapped observed image. Scoring measures the degree of a match between a character template and the observed image. Computation of these actual scores is replaced with the simpler computation of column-based (i.e., one-dimensional) heuristic scores. Because the column-based heuristic scores can be shown to be a true upper bound on actual template-image scores, the heuristic scores are accurate enough to use in place of actual scoring during text line decoding. The heuristic scores essentially reduce the expensive two-dimensional computation of the actual template-image scores required by prior decoding methods to a simpler but accurate one-dimensional computation.

Document image decoding using text line column-based heuristic scoring

A text recognition system represents the decoded message of a document image as a path through an image network. A method for integrating a language model into the network selectively expands the network to accommodate the language model only for certain ones of the paths in the network, effectively managing the memory storage requirements and computational complexities of integrating the language model efficiently into the network. The language model generates probability distributions indicating the probability of a certain character occurring in a string, given one or more previous characters in the string. Selectively expanding the image network is achieved by initially using upper bounds on the language model probabilities on the branches of an unexpanded image network. A best path search operation is then performed to determine an estimated best path through the image network using these upper bound scores. After decoding, only the nodes on the estimated best path are expanded with new nodes and with branches incoming to the new nodes that accommodate new language model scores reflecting actual character histories in place of the upper bound scores. Decoding and selectively expanding the image network are repeated until the final output transcription of the text image has been produced.

Document image decoding using an integrated stochastic language model

A system for printing glyph frames around known obstructions. All frames in an area are determined to be obstructed or unobstructed, based on their location with respect to other printed areas. The unobstructed locations can be numbered and glyph data printed within. In the alternative, the good locations can be numbered modulo some number much smaller that the number of available locations to provide redundancy. The unobstructed locations can be stored in either the sync lines or in the data area of other locations known to be unobstructed. Also, the frame itself can be identified as obstructed or unobstructed to provide more redundancy.

System for embedded digital data that allows embedding of data around known obstructions

Growing community of inventors

Palo Alto, CA, United States of America

Dan S Bloomberg