Image analysis – Applications – Reading aids for the visually impaired
Reexamination Certificate
1998-12-30
2001-02-13
Johns, Andrew W. (Department: 2721)
Image analysis
Applications
Reading aids for the visually impaired
Reexamination Certificate
active
06188779
ABSTRACT:
BACKGROUND
This invention relates generally to reading machines which are used as part of a remedial reading program to assist individuals with learning disabilities or severe visual impairments or blindness.
Reading machines have been used to improve the educational attainment of individuals with learning disabilities or severe visual impairments or blindness. In general, known reading machines are computer based. Reading machines include specialized software to process an input source document and generate synthetic speech to enable a user to hear the computer read through the document a word, line, sentence, etc. at a time. Often these reading machines include a scanner to provide one technique to input source documents to the reading machine.
Common types of source materials that are scanned into a reading machine include magazines and books. Often the books and/or magazines are arranged on the scanner in a manner such that two pages of the book are scanned in a common scan producing an image file with two pages. When optical character recognition software processes the image file produced from scanning such a book, the optical character recognition software can mistake the two pages of the book for a pair of columns on a single page. This especially becomes a problem when the page itself contain columns.
SUMMARY
According to an aspect of the invention, a method of splitting an image file produced by scanning a document includes determining probable regions of text from an image file that contains an image of a left page and a right page of a document and determining a page boundary between the regions of text. The method also includes splitting the image file about the determined page boundary into two separate image files corresponding to an image of the left hand page and the right hand page.
According to a further aspect of the invention, a computer program product resides on a computer readable medium. The computer program product includes instructions for separating an image file containing a pair of pages into separate image files for each page. The computer program includes instructions for causing a computer to determine probable regions of text of an image file and determine from the probable regions of text a page boundary between the regions of text. The program also includes instructions to cause a computer to split the image file into two separate image files corresponding to an image of the left hand page of the image file and the right hand page of the image file text.
According to a further aspect of the invention, a method of pre-processing an image file for optical character recognition includes separating the image file containing images of a pair of pages into two separate files each containing an image of a single page and deskewing each of the separated image files. The method further includes applying each of the deskewed image files to an optical character recognition software.
According to a still further aspect of the invention, a method of operating a reading system includes scanning a source document to produce an image file containing an image of a left page and a right page of a document. The method also includes determining probable regions of text of the image file, determining from the probable regions of text a page boundary between the regions of text, and splitting the image file into two separate image files corresponding to an image of the left hand page of the image file and the right hand page of the image file.
According to a still further aspect of the invention, a computer program product resides on a computer readable medium. The computer program product separates an image file containing a pair of pages into separate image files of each page. The product also synthesizes speech for the separated image files. The program includes instructions for causing a computer to scan a source document to produce an image file containing an image of a left page and a right page of a document and determine probable regions of text of the image file. The program also includes instructions to determine from the probable regions of text a page boundary between the regions of text and split the image file into two separate image files corresponding to an image of the left hand page of the image file and the right hand page of the image file. The program also uses optical character recognition to separately recognize the text from each image file in order to synthesize speech from each separate page of text.
According to a still further aspect of the invention, a method of performing optical character recognition on an image file comprised of a left side page and a right side page includes determining probable regions of text from the image file and determining a page boundary between the regions of text. The method splits the image file about the determined page boundary into two separate image files corresponding to an image of the left hand page and the right hand page and converts the two separate image files into two blocks of text corresponding to the left page and the right page.
One or more of the following advantages may be provided by one or more aspects of the invention. Often a user places a document on a scanner in such a manner that the scanner scans two pages of a document into a single image file. The various aspects of the invention process the image file to produce two separate image files. Optical character recognition is used to separately recognize the text from each image file. If text files produced from the optical character recognition are sent to a speech synthesizer the portions of the document corresponding to the separate files are read in the correct order. The processing of the image file will insure that text from the right hand page is read after the text from the left hand page.
Often an image file that contains an image of two pages of a document may have text on the page skewed with respect to the text on the other page. It is desirable for improving optical character recognition to deskew the image files so that the text within the image file is perpendicular to the boundary of the page. The composite image file is separated into two separate image files and fed to a deskew filter to produce a deskewed image file. This deskew preprocessing can significantly improve optical character recognition performance. In systems that use the recognized text, it is often useful to have the text organized in pages that correspond to the pages in the original document.
REFERENCES:
patent: 5774580 (1998-06-01), Saitoh
Bromberg & Sunstein LLP
Johns Andrew W.
L&H Applications USA, Inc.
Nakhajvan Shervin
LandOfFree
Dual page mode detection does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Dual page mode detection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dual page mode detection will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2564843