Optical character recognition using baseline information

• US 4,251,799 A
• Filed: 03/30/1979
• Issued: 02/17/1981
• Est. Priority Date: 03/30/1979
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Jih, Chentung R.
• Primary Examiner(s): Boudreau, Leo H.
• Application Number:
  US06/025,436
• Time in Patent Office: 690 Days
• Field of Search: 340/146.3 H, 340/146.3 Y, 340/146.3 SY, 340/146.3 SG
• US Class Current: 382/174
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1478 of characters or characters lines

Abstract

An Optical Character Recognition (OCR) system for printed characters uses information about baseline location to assist in registration, segmentation, recognition and the separation of underscoring. Selected ones of the sensor elements in the OCR scanning array are identified as defining a baseline searching window and the outputs of these selected elements are utilized to detect baseline information for each character or group of characters in a printed line.

Claims

1. A method for processing image data from a scanning array which scans characters in a scanning direction to locate a baseline, comprising the steps of

calculating a horizontal projection histogram of said characters at a plurality of periodic spaced locations along said scanning direction,

detecting a peak characteristic of each of said histograms, and

averaging said detected peak characterstics of said histograms to produce an estimate of the skew of said baseline with respect to said scanning direction.

2. The method in accordance with claim 1 including the step of utilizing said estimated baseline data to dynamically alter said scanning array so that said scanning array follows said estimated baseline to process the character image data.

3. The method in accordance with claim 1, in which a histogram is calculated for each scanned character.