DETECTING TEXT USING STROKE WIDTH BASED TEXT DETECTION

• US 20090285482A1
• Filed: 05/19/2008
• Published: 11/19/2009
• Est. Priority Date: 05/19/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ofek, Eyal, Wexler, Yonatan, Epshtein, Boris
• Granted Patent: US 8,917,935 B2
• Time in Patent Office: Days
• US Class Current: 382/176
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/414 Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text

Abstract

Detecting text using stroke width based text detection. As a part of the text detection, a representation of an image is generated that includes pixels that are associated with the stroke widths of components of the image. Connected components of the image are identified by filtering out portions of the pixels using metrics related to stroke width. Text is detected in the image based on the identified connected components.

Claims

1. A method for detecting text using stroke width based text detection, comprising:

generating a representation of an image that includes pixels that are associated with stroke widths of components of said image;

identifying connected components of said image by filtering out portions of said pixels using metrics related to stroke width to exclude said portions of said pixels; and

based on said identifying said connected components, detecting text in said image.

2. The method of claim 1 wherein said image is selected from the group that includes natural image, infra red image, CT scan and radar.

3. The method of claim 1 wherein said image is based on vector data.

4. The method of claim 1 wherein said image is presented on a non-regular grid.

5. The method of claim 1 wherein said filtering comprises eliminating components that cannot be recognized or are not considered associated with valid text.

6. The method of claim 1 wherein said filtering comprises determining whether adjacent pixels are associated with a similar stroke width.

7. The method of claim 1 wherein said filtering comprises eliminating components with an associated stroke width whose standard deviation exceeds a predetermined threshold.

8. A method for associating pixels with the width of image components for detecting text in images, comprising:

determining image components of said image that are characterized by opposing elements;

projecting lines between said opposing elements on both sides of a stroke;

associating pixels along said lines with widths of particular image components;

identifying connected components of said image by filtering out portions of said pixels using metrics related to the widths of said particular image components to exclude said portions of said pixels; and

based on said identifying said connected components, detecting text in said image.

9. The method of claim 8 wherein said opposing elements on both sides of a stroke are represented by edges in said image.

10. The method of claim 8 wherein said image is selected from the group consisting of natural image, infra red image, CT scan, radar, depth image, MRI, X-Ray and UltraSound.

11. The method of claim 8 wherein said image is based on vector data.

12. The method of claim 8 wherein said image is presented on a non-regular grid.

13. The method of claim 8 wherein said filtering comprises eliminating components that cannot be recognized or that are not considered to be associated with valid text.

14. The method of claim 8 wherein said filtering comprises determining whether adjacent pixels are associated with a similar stroke width.

15. The method of claim 8 wherein said filtering comprises eliminating components with an associated stroke width whose standard deviation exceeds a predetermined threshold.

16. A computer-readable medium having computer-executable module, comprising:

a stroke width data generating module for generating a representation of an image that includes elements that are associated with the stroke widths of components of said image;

a connected component identifying module for identifying connected components of said image by filtering out portions of said elements using metrics related to stroke width to exclude said portions of said elements; and

a text detecting module for detecting text in said image based on said identifying said connected components.

17. The medium of claim 16 wherein said image is selected from the group that includes natural image, infra red image, CT scan, depth image, Ultra Sound, MRI, and radar.

18. The medium of claim 16 wherein said image is based on vector data.

19. The medium of claim 16 wherein said image is presented on a non-regular grid.

20. The medium of claim 16 wherein said filtering comprises determining whether adjacent elements are associated with a similar stroke width.