Methods and systems for identifying text in digital images

• US 7,876,959 B2
• Filed: 09/06/2006
• Issued: 01/25/2011
• Est. Priority Date: 09/06/2006
• Status: Expired due to Fees

• Current Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Original Assignee: Sharp Laboratories of America Incorporated (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Matsuda, Toyohisa, Campbell, Richard John, Chen, Lawrence Shao-hsien
• Primary Examiner(s): Dang; Duy M
• Application Number:
  US11/470,519
• Publication Number:
  US 20080056573A1
• Time in Patent Office: 1,602 Days
• Field of Search: 382/173, 382/176, 382/286, 382/302, 358/3.01
• US Class Current: 382/176
• CPC Class Codes: G06V 30/413 Classification of content, e.g. text, photographs or tables

Abstract

Aspects of the present invention relate to systems, methods and devices for detection of text in an image using an initial text classification result and a verification process. In particular, a support region of a candidate text pixel in a text-candidate map may be expanded to produce a revised text-candidate map. Pictorial regions in the image may be discriminated based on an entropy measure using masking and the revised text-candidate map, and the revised text-candidate map may be refined based on the pictorial regions.

Claims

1. A method for identifying text in a digital image, said method comprising:

a) in an image processing system comprising at least one computing device, expanding a support region of a candidate text pixel in a text-candidate map, wherein said expanding comprises;

i) receiving an edge map wherein said edge map identifies edges in said digital image;

ii) generating an edge count wherein said edge count generating comprises associating an entry in said edge count with a neighborhood in said edge map and the value of said entry in said edge count is the sum of the edge pixels in said neighborhood in said edge map;

iii) receiving a text-candidate map wherein said text-candidate map identifies text-candidate pixels in said digital image; and

iv) generating a text count wherein said text count generating comprises an entry in said text count with a neighborhood in said text-candidate map and the value of said entry in said text count is the sum of the text-candidate pixels in said neighborhood in said text-candidate map,

thereby producing a revised text-candidate map;

b) in said image processing system, discriminating pictorial regions in said digital image based on an entropy measure comprising masking using said revised text-candidate map; and

c) in said image processing system, refining said revised text-candidate map based on said pictorial regions.

2. A method as described in claim 1 wherein said discriminating further comprises averaging said entropy measure.

3. A method as described in claim 1 wherein said discriminating further comprises hysteresis thresholding.

4. A method as described in claim 1 wherein said candidate-text map is based on a texture feature of said digital image.

5. A system for identifying text in a digital image, said system comprising:

a) an expander processor for expanding a support region of a candidate text pixel in a text-candidate map, wherein said expander processor comprises;

i) an edge map receiver for receiving an edge map wherein said edge map identifies edges in said digital image;

ii) an edge count generator for generating an edge count wherein said edge count generating comprises associating an entry in said edge count with a neighborhood in said edge map and the value of said entry in said edge count is the sum of the edge pixels in said neighborhood in said edge map;

iii) a text-candidate map receiver for receiving a text-candidate map wherein said text-candidate map identifies text-candidate pixels in said digital image; and

iv) a text count generator for generating a text count wherein said text count generating comprises associating an entry in said text count with a neighborhood in said text-candidate map and the value of said entry in said text count is the sum of the text-candidate pixels in said neighborhood in said text-candidate map,

thereby producing a revised text-candidate map;

b) a discriminator processor for discriminating pictorial regions in said digital image based on an entropy measure comprising masking using said revised text-candidate map; and

c) a refiner processor for refining said revised text-candidate map based on said pictorial regions.

6. A system as described in claim 5 wherein said discriminating further comprises averaging said entropy measure.

7. A system as described in claim 5 wherein said discriminating further comprises hysteresis thresholding.

8. A system as described in claim 5 wherein said candidate-text map is based on a texture feature of said digital image.

9. A system as described in claim 5 wherein said discriminating further comprises a plurality of scan passes.