Page segmentor

• US 4,817,169 A
• Filed: 04/14/1987
• Issued: 03/28/1989
• Est. Priority Date: 04/14/1987
• Status: Expired due to Fees

• Current Assignee: Nippon Sheet Glass Company Limited
• Original Assignee: Nippon Sheet Glass Company Limited
• Inventors: Ueno, Hiroshi, Young, James R., Peppers, Norman A., Nishi, Hisami
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): NOT, DEFINED
• Application Number:
  US07/038,631
• Time in Patent Office: 714 Days
• Field of Search: 382/9, 382/32, 382/44, 382/45, 382/35, 356/394
• US Class Current: 382/173
• CPC Class Codes: G06V 10/88 Image or video recognition using optical means, e.g. reference filters, holographic masks, frequency domain filters or spatial domain filters

Abstract

In a page segmentor for dividing a page into areas each having a single feature, the division being performed as preprocessing for pattern recognition, one of the areas within the page is multiplied into multiple images, a feature of each area is extracted according to the presence or absence of each of pairs between the multiple images and a plurality of different masks, and the page is divided into areas until each area has a single feature. High-speed page segmentation can be achieved so as to obtain a desired feature.

Claims

1. A page segmentor comprising:

dividing means for dividing a page into a plurality of areas,

multiplying means for multiplying an image of one of the plurality of areas into a plurality of identical, smaller images, said multiplying means including a single convex lens having a relatively large aperture size, the image of the area being located at a focal point of said single convex lens, and a plurality of convex lenses each having a relatively small aperture size, said plurality of convex lenses being aligned on a side of said single convex lens opposite the image of the area in a direction perpendicular to an optical axis of said single convex lens in an optical path of light passing from the image of the area through said single convex lens,

feature extracting means for simultaneously extracting features of the area by optically detecting the presence or absence of a correlation between each of pairs of the plurality of smaller images and a plurality of different masks, said plurality of masks including at least one mask for detecting a laterally written document having character strings written horizontally across the page, and at least one other mask for detecting a vertically written document having character strings written up or down the page, and

merging means for merging two or more of the areas which have identical single features,

wherein the page is further divided by the dividing means into progressively smaller areas until each of the areas has a single feature.

2. A segmentor according to claim 1, wherein said plurality of masks further include a mask having slits inclined at a given angle.

3. A segmentor according to claim 2, wherein said plurality of masks further include a mask having a shape corresponding to the shape of a given figure or a given table.