US4635290A — Sectioning apparatus and method for optical chara…

Claim 1

Independent Claim 160 words

A character sectioning apparatus for segmenting character string images into individual characters by means of a projection distribution obtained through a series of character string images, comprising: a means for identifying a plurality of character lumps, separated by spaces, from said projection distribution; means for estimating a character pitch; means means responsive to said projection distribution and said character pitch for setting segmentation candidate sections comprising a plurality of segmentation candidate positions; means for calculating a distance measurement between each of said segmentation candidate positions within said segmentation candidate sections and segmentation candidate positions within an adjacent segmentation candidate section; means for calculating an appreciation standard as a function of said distance measurements for segmentation candidate positions in each of said plurality of segmentation candidate sections; and means for selecting an optimum segmentation candidate position for each of said plurality of segmentation candidate sections; wherein each optimum segmentation candidate position minimizes said appreciation standard for the respective segmentation candidate section.

Claim 2

Dependent Claim 160 words

A character sectioning apparatus for segmenting character string images into individual characters by means of a projection distribution obtained through a series of character string images, comprising: a means for identifying a plurality of character lumps, separated by spaces, from said projection distribution; means for estimating a character pitch; means means responsive to said projection distribution and said character pitch for setting segmentation candidate sections comprising a plurality of segmentation candidate positions; means for calculating a distance measurement between each of said segmentation candidate positions within said segmentation candidate sections and segmentation candidate positions within an adjacent segmentation candidate section; means for calculating an appreciation standard as a function of said distance measurements for segmentation candidate positions in each of said plurality of segmentation candidate sections; and means for selecting an optimum segmentation candidate position for each of said plurality of segmentation candidate sections; wherein each optimum segmentation candidate position minimizes said appreciation standard for the respective segmentation candidate section.

Claim 3

Dependent Claim 33 words

The character sectioning apparatus as claimed in claim 1, wherein said segmentation candidate section is a logical product of said blank space and a section set as a function of said character pitch.

Claim 4

Dependent Claim 90 words

The character sectioning apparatus as claimed in claim 1, further comprising means for identifying a character lump which exceeds a first threshold and setting a segmentation candidate section defined by the logical sum of a first and second logical product, wherein: said first logical product is a product of the amount of said character lump exceeding said first threshold and a section set as a function of said character pitch; and said second logical product is a product of a blank space and a section set by said character pitch.

Claim 5

Dependent Claim 31 words

The apparatus of claim 3, wherein said character lump which exceeds a first threshold is truncated at a leading end and a trailing end by amount equal to a fifth threshold.

Claim 6

Dependent Claim 16 words

The apparatus of claim 4, wherein said fifth threshold is a function of said character pitch.

Claim 7

Dependent Claim 57 words

The character sectioning apparatus as claimed in claim 1, further comprising means for comparing the width of said blank spaces with a second threshold and for extracting a blank space having a width greater than said second threshold, wherein said extracted blank space is set as a first trailing end candidate section of said character string image.

Claim 8

Dependent Claim 113 words

The character sectioning apparatus of claim 6, further comprising: means for extracting a first and a second section from said extracted blank space having a width greater than said second threshold; wherein said first section extends from a leading end of said extracted blank space to a position determined as a function of said character pitch and a third threshold; and wherein said second section extends from a leading end of a character lump preceding said extracted blank space to a position determined as a function of said character pitch and a fourth threshold; and wherein a second trailing end section is defined as the logical sum of said first and second sections.

Claim 9

Dependent Claim 30 words

The apparatus of claim 1, wherein said appreciation standard is a function of a mean value of said distance measurements, a variance of said distance measurements, and said character pitch.

Claim 10

Dependent Claim 175 words

A character sectioning apparatus for segmenting character string images into individual characters by scanning a series of character string images, comprising: means for detecting a plurality of character lumps from a series of character string images; means for calculating a distance measurement between each of said character lumps and storing a frequency distribution of said distance measurements; means for forecasting a presence section comprising a plurality of candidate character pitches; means for dividing said frequency distribution into a plurality of areas; means for detecting a candidate character pitch in said presence section, said candidate pitch minimizing an estimation error appreciation standard in a plurality of said areas, wherein said estimation error appreciation standard is a function of an optimum linear sum of estimation errors of the character pitch; means for calculating a variance of said distance measurements and for calculating a variance appreciated value said variance appreciated value being a function of an optimum linear sum of a plurality of said variances; and means for identifying the character pitch according to said variance appreciated value.

Claim 11

Dependent Claim 175 words

A character sectioning apparatus for segmenting character string images into individual characters by scanning a series of character string images, comprising: means for detecting a plurality of character lumps from a series of character string images; means for calculating a distance measurement between each of said character lumps and storing a frequency distribution of said distance measurements; means for forecasting a presence section comprising a plurality of candidate character pitches; means for dividing said frequency distribution into a plurality of areas; means for detecting a candidate character pitch in said presence section, said candidate pitch minimizing an estimation error appreciation standard in a plurality of said areas, wherein said estimation error appreciation standard is a function of an optimum linear sum of estimation errors of the character pitch; means for calculating a variance of said distance measurements and for calculating a variance appreciated value said variance appreciated value being a function of an optimum linear sum of a plurality of said variances; and means for identifying the character pitch according to said variance appreciated value.

Claim 12

Dependent Claim 173 words

A character pitch detector for detecting a character pitch of a series of character string images, comprising: means for detecting a plurality of character lumps from a series of said character string images; means for calculating a distance measurement between said character lumps and for storing a frequency distribution of said distance measurements; means for forecasting a presence section comprising candidate character pitches; means for dividing said frequency distribution into a plurality of areas by means of an arbitrary candidate character pitch included in said presence section; means for calculating a character pitch estimation error between a representative point of a plurality of said distance measurements in each of said areas and said arbitrary candidate character pitch; means for calculating an estimation error appreciation standard as a function of an optimum linear sum of said character pitch estimation error in a plurality of said areas; and means for calculating an optimum character pitch, wherein said optimum character pitch mimimizes said estimation error appreciation standard within a plurality of said arbitrary candidate character pitches.

Claim 13

Dependent Claim 173 words

A character pitch detector for detecting a character pitch of a series of character string images, comprising: means for detecting a plurality of character lumps from a series of said character string images; means for calculating a distance measurement between said character lumps and for storing a frequency distribution of said distance measurements; means for forecasting a presence section comprising candidate character pitches; means for dividing said frequency distribution into a plurality of areas by means of an arbitrary candidate character pitch included in said presence section; means for calculating a character pitch estimation error between a representative point of a plurality of said distance measurements in each of said areas and said arbitrary candidate character pitch; means for calculating an estimation error appreciation standard as a function of an optimum linear sum of said character pitch estimation error in a plurality of said areas; and means for calculating an optimum character pitch, wherein said optimum character pitch mimimizes said estimation error appreciation standard within a plurality of said arbitrary candidate character pitches.

Claim 14

Dependent Claim 159 words

A character pitch detecting method for detecting a character pitch of a series of character string images, comprising the steps of: detecting a plurality of character lumps from a series of character string images; calculating distance measurements between said character lumps; storing a frequency distribution of said distance measurements; forecasting a presence section comprising candidate character pitches; dividing said frequency distribution into a plurality of areas by means of an arbitrary candidate character pitch contained in said presence section; calculating a character pitch estimation error between a representative point of a plurality of said distance measurements in each of said areas and said arbitrary candidate character pitch; calculating an estimation error appreciation standard as a function of an optimum linear sum of said character pitch estimation errors in a plurality of said areas; and calculating an optimum character pitch wherein said optimum character pitch minimizes said estimation error appreciation standard within a plurality of said arbitrary candidate character pitches.

Claim 15

Dependent Claim 159 words

A character pitch detecting method for detecting a character pitch of a series of character string images, comprising the steps of: detecting a plurality of character lumps from a series of character string images; calculating distance measurements between said character lumps; storing a frequency distribution of said distance measurements; forecasting a presence section comprising candidate character pitches; dividing said frequency distribution into a plurality of areas by means of an arbitrary candidate character pitch contained in said presence section; calculating a character pitch estimation error between a representative point of a plurality of said distance measurements in each of said areas and said arbitrary candidate character pitch; calculating an estimation error appreciation standard as a function of an optimum linear sum of said character pitch estimation errors in a plurality of said areas; and calculating an optimum character pitch wherein said optimum character pitch minimizes said estimation error appreciation standard within a plurality of said arbitrary candidate character pitches.

Claim 16

Independent Claim 141 words

A method of segmenting a character string image into individual characters comprising the steps of: (a) identifying character lumps and blank spaces within said character string; (b) estimating a character pitch; (c) defining segmentation candidate sections comprising segmentation candidate positions based on said character pitch; (d) measuring a distance between each segmentation candidate position within a first segmentation candidate section and a segmentation candidate position within a second segmentation candidate section adjacent to said first segmentation candidate section; (e) calculating an appreciation standard as a function of said distance measurements for a segmentation candidate position in each segmentation candidate section; (f) selecting an optimum segmentation candidate position for each segmentation candidate section, wherein said optimum segmentation candidate position minimizes said appreciation standard for the respective segmentation candidate section; and (g) segmenting said character string image at said optimum segmentation candidate positions.

Claim 17

Independent Claim 141 words

A method of segmenting a character string image into individual characters comprising the steps of: (a) identifying character lumps and blank spaces within said character string; (b) estimating a character pitch; (c) defining segmentation candidate sections comprising segmentation candidate positions based on said character pitch; (d) measuring a distance between each segmentation candidate position within a first segmentation candidate section and a segmentation candidate position within a second segmentation candidate section adjacent to said first segmentation candidate section; (e) calculating an appreciation standard as a function of said distance measurements for a segmentation candidate position in each segmentation candidate section; (f) selecting an optimum segmentation candidate position for each segmentation candidate section, wherein said optimum segmentation candidate position minimizes said appreciation standard for the respective segmentation candidate section; and (g) segmenting said character string image at said optimum segmentation candidate positions.

Claim 18

Dependent Claim 65 words

The method of claim 12, wherein the step of defining segmentation candidate sections comprises the steps of: (h) defining a first section as a function of said character pitch; (i) calculating a first logical product of said first section and a blank space; (j) defining a segmentation candidate by said first logical product; and (k) repeating steps h-j over the length of the character string.

Claim 19

Dependent Claim 89 words

The method of claim 12, wherein the step of defining a segmentation candidate section comprises: (h) identifying character lumps which exceed a first threshold; (i) calculating a first logical product of a section set as a function of said character pitch and the amount of said character lump exceeding said first threshold; (j) calculating a second logical product of a blank space and a section set by said character pitch; and (k) setting a segmentation candidate section equal to a logical sum of said first and second logical products.

Claim 20

Dependent Claim 36 words

The method of claim 14, further comprising the step of: (j) truncating said character pumps which exceed a first threshold at a leading end and a trailing end by an amount equal to a fifth threshold.

Claim 21

Dependent Claim 43 words

The method of claim 12 further comprising the steps of: (h) identifying a first blank space which exceeds a second threshold; (i) setting said first blank space which exceeds said second threshold as a first trailing end candidate section of said character string.

Claim 22

Dependent Claim 110 words

The method of claim 16 further comprising the steps of: (j) calculating a third threshold as a function of said character pitch; (k) calculating a fourth threshold as a function of said character pitch; (l) defining a first trailing end permissible section extending from a leading end of said first blank space to said third threshold; (m) defining a second trailing end permissible section extending from a leading end of a character lump immediately preceding said first blank space to said fourth threshold; (n) calculating a logical sum of said first and second trailing end permissible sections; and (o) defining said logical sum as a second trailing end candidate section.

Claim 23

Independent Claim 170 words

A method of segmenting character strings into individual characters comprising the steps of: (a) detecting a plurality of character lumps from a series of character string images; (b) calculating a distance measurement between each of said character lumps; (c) storing a frequency distribution of said distance meaurements; (d) forecasting a presence section comprising a plurality of candidate character pitches; (e) dividing said frequency distribution into a plurality of areas; (f) detecting a candidate character pitch in said presence section, wherein said candidate character pitch minimizes an estimation error appreciation standard in a plurality of said areas, said estimation error appreciation standard being a function of an optimum linear sum of estimation errors of the character pitch; (g) calculating a variance of said distance measurements; (h) calculating a variance appreciated value, said variance appreciated value comprising an optimum linear sum of a plurality of said variances; (i) identifying the character pitch according to said variance appreciated value; and (j) segmenting the character string as a function of the identified character pitch.

Claim 24

Independent Claim 170 words

A method of segmenting character strings into individual characters comprising the steps of: (a) detecting a plurality of character lumps from a series of character string images; (b) calculating a distance measurement between each of said character lumps; (c) storing a frequency distribution of said distance meaurements; (d) forecasting a presence section comprising a plurality of candidate character pitches; (e) dividing said frequency distribution into a plurality of areas; (f) detecting a candidate character pitch in said presence section, wherein said candidate character pitch minimizes an estimation error appreciation standard in a plurality of said areas, said estimation error appreciation standard being a function of an optimum linear sum of estimation errors of the character pitch; (g) calculating a variance of said distance measurements; (h) calculating a variance appreciated value, said variance appreciated value comprising an optimum linear sum of a plurality of said variances; (i) identifying the character pitch according to said variance appreciated value; and (j) segmenting the character string as a function of the identified character pitch.

Annuity Stage	Window Opens	Payment Deadline	Surcharge Period Until	Estimated Fee	Status
3.5 Years (Year 4)	1990-01-05	1990-07-07	1991-01-06	$2000.00	Window Closed / Lapsed
7.5 Years (Year 8)	1994-01-05	1994-07-07	1995-01-06	$3760.00	Window Closed / Lapsed
11.5 Years (Year 12)	1998-01-05	1998-07-07	1999-01-06	$7700.00	Window Closed / Lapsed

Sectioning apparatus and method for optical character reader systems

Abstract

Inventors

Bibliographic Details

Claims Analysis

References Cited (Backward Citations)

USPTO Annuity & Maintenance Windows

Assigned Categories & Tags

Assign Category Tag