CHARACTER LINE RECOGNITION METHOD AND CHARACTER LINE RECOGNITION DEVICE

• US 20100135578A1
• Filed: 11/23/2009
• Published: 06/03/2010
• Est. Priority Date: 11/23/2008
• Status: Active Grant

• Current Assignee: Nidec Sankyo Corporation (Nidec Corporation)
• Original Assignee: Nidec Sankyo Corporation (Nidec Corporation)
• Inventors: NAKAMURA, Hiroshi
• Granted Patent: US 8,401,299 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06V 10/10 Image acquisition document image scanning and transmission H04N1/00; control of digital cameras H04N23/60
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/162 Quantising the image signal
• CPC Class Codes: G06V 30/18095 Summing image-intensity values; Projection and histogram analysis

Abstract

A character line recognition method for processing image data obtained by scanning a character line on a medium to recognize the character line may include processing the image data into monochrome binary format image data by using a predetermined binarization standard threshold; extracting character features from each character that composes the character line to calculate similarity with respect to standard character features; temporarily determining characters based on the similarity; calculating basic statistics of the similarity for all the characters which have been temporarily determined; and changing the binarization standard threshold based on the basic statistics and then returning to the processing the image data.

Claims

1. A character line recognition method for processing image data obtained by scanning a character line on a medium to recognize the character line, the method comprising:

processing the image data into monochrome binary format image data by using a predetermined binarization standard threshold;

extracting character features from each character that composes the character line to calculate similarity with respect to standard character features;

temporarily determining characters based on the similarity;

calculating basic statistics of the similarity for all the characters which have been temporarily determined; and

changing the binarization standard threshold based on the basic statistics and then returning to the processing the image data.

2. The character line recognition method as set forth in claim 1 further comprising, between the end of the processing the image data and the extracting character features:

calculating an orthogonal projection of the binary format image data in a direction orthogonal to a direction of the character line while shifting along the binary format image data in the direction of the character line;

detecting positions at which orthogonal projection data obtained in the calculating an orthogonal projection are above a predetermined boundary judgment threshold as segmenting-positions for segmenting the characters that compose the character line; and

obtaining rectangles bounding the characters based on the segmenting-positions detected in the detecting positions;

wherein the extracting character features extracts character features in the rectangles and calculates similarity with respect to standard character features.

3. The character line recognition method asset forth in claim 1, wherein the changing the binarization standard threshold classifies a recognition result into one of a plurality of evaluation categories based on the basic statistics, and changes the binarization standard threshold based on the evaluation category in which the recognition result is classified.

4. The character line recognition method asset forth in claim 3 wherein the changing the binarization standard threshold, having a minimum value of the similarity as the basic statistics, classifies the recognition result into one of the evaluation categories according to the minimum value.

5. The character line recognition method asset forth in claim 1, wherein the changing the binarization standard threshold changes the binarization standard threshold so as to improve the basic statistics.

10. A character line recognition device equipped with a character line recognition means to recognize a character line on a medium by using the character line recognition method asset forth in claim 1.

6. A character line recognition method for processing image data obtained by scanning a character line on a medium to recognize the character line, the method comprising:

processing the image data into monochrome binary format image data by using a predetermined binarization standard threshold;

extracting character features from each character composing the character line to calculate similarity with respect to standard character features;

temporarily determining characters based on the similarity;

classifying a recognition result of the entire character line, which includes all the characters temporarily determined, into one of a plurality of evaluation categories; and

changing the binarization standard threshold based on the evaluation category in which the recognition result is classified and then returning to the processing the image data.

7. The character line recognition method as set forth in claim 6, wherein the changing the binarization standard threshold changes the binarization standard threshold to be used in re-processing, based on results from the repeated extracting character features and temporarily determining characters, so as to improve the recognition result of the entire character line.

8. The character line recognition method asset forth in claim 6, further comprising:

calculating basic statistics of the similarity on all the characters which have been temporarily determined;

wherein the changing the binarization standard threshold changes the binarization standard threshold based on the evaluation category in which a result is classified in the classifying the recognition result and on the basic statistics calculated.

9. The character line recognition method asset forth in claim 8 wherein the changing the binarization standard threshold changes the binarization standard threshold so as to improve the basic statistics.

11. A character line recognition device equipped with a character line recognition means to recognize a character line on a medium by using the character line recognition method asset forth in claim 6.