Automated method for extracting highlighted regions in scanned source
View Patent ↗An automated method for extracting highlighted regions in a scanned text documents includes color masking of highlight regions, extracting text from highlighted regions, recognizing the characters in extracted text optically and inserting the recognized characters to new document in order to easily identify highlighted text in scanned images. Using a two-layer multi-mask compression technology configured in a scanned export image path, edges and text regions can be extracted and together with the use of mask coordinates and associated mask colors, all highlighted texts can be easily identified and extracted. Optical Character Recognition (OCR) can then be utilized to appropriate summarization of different extracted highlighted texts.
1. A method for extracting highlighted regions in a scanned text document, comprising:
converting a color space of a scanned text document comprising a highlighted text within a highlighted region into an sRGB color space by converting said scanned text document into a plurality of background image planes and a plurality of mask binary planes;
applying constant threshold intervals for each color channel of said sRGB color space;
extracting said highlighted region from said scanned text document by bounding a color of said highlighted region to a highlighted cuboid in said sRGB color space by determining if each background image plane of said plurality of background image planes is a uniform color;
extracting said highlighted text from said highlighted cuboid by creating a highlighted region image by setting pixel values of uniform color pixels within said highlighted cuboid to white and remaining pixels within said highlighted region to black; and
optically recognizing said highlighted text in order to recognize text extracted from said highlighted region image of said scanned text document.
2. The method of claim 1 , further comprising:
generating recognized text, in response to optically recognizing said highlighted text; and
inserting said recognized text into a new document.
3. The method of claim 2 , further comprising:
summarizing a collection of scanned original text using an optically recognized highlighted text within said scanned text document to generate a summarized document.
4. The method of claim 3 further comprising rendering said summarized document utilizing a printer.
5. The method of claim 3 wherein said summarized document is rendered remotely.
6. The method of claim 5 wherein said summarized document is rendered remotely utilizing an electronic facsimile transmission.
7. The method of claim 5 wherein said summarized document is rendered remotely utilizing electronic mail.
8. The method of claim 3 further comprising electronically storing said summarized within an electronic repository.
9. The method of claim 1 further comprising:
extracting said highlighted text from said highlighted cuboid utilizing a 4-connected seed fill operation.
10. A method for extracting highlighted regions in a scanned image document, comprising:
converting a color space of a scanned image document into an sRGB color space;
converting said sRGB color space into a plurality of background image planes and a plurality of mask binary regions;
analyzing said plurality of background image planes utilizing at least one mask coordinate, wherein said plurality of background image planes are located beneath at least one mask binary region among said plurality of mask binary regions; and
optically recognizing a highlighted text in said at least one mask binary region, by setting pixel values of uniform color pixels within said at least one mask binary region to white and remaining pixels within said at least one mask binary region to black if said plurality of background image planes beneath said at least one mask binary region comprises a uniform color.
11. A system for extracting highlighted regions in a scanned text document, comprising:
a data-processing apparatus comprising processor readable code embodied on one or more processor readable storage devices, said processor readable code comprising:
a module for converting a color space of a scanned text document comprising a highlighted text within a highlighted region into an sRGB color space by converting said scanned text document into a plurality of background image planes and a plurality of mask binary planes;
a module for extracting said highlighted region from said scanned text document by bounding a color of said highlighted region to a highlighted cuboid in said sRGB color space by determining if each background image plane of said plurality of background image planes is a uniform color; and
a module for optically recognizing said highlighted text in order to recognize text extracted from said highlighted region of said scanned text document by creating a highlighted region image by setting pixel values of uniform color pixels within said highlighted cuboid to white and remaining pixels within said highlighted region to black.
12. The system of claim 11 , further comprising:
a module for generating recognized text, in response to optically recognizing said highlighted text; and
a module for inserting said recognized text into a new document.
13. The system of claim 12 , further comprising a collection of scanned original text summarized using an optically recognized highlighted text within said scanned text document to generate a summarized document.
14. The system of claim 13 wherein said summarized document is rendered utilizing a printer.
15. The system of claim 13 wherein said summarized document is rendered remotely.
16. The system of claim 15 wherein said summarized document is rendered remotely utilizing an electronic facsimile transmission.
17. The system of claim 15 wherein said summarized document is rendered remotely utilizing electronic mail.
18. The system of claim 15 wherein said summarized document is electronically stored within an electronic repository.
19. The method of claim 10 further comprising:
extracting said highlighted text from said at least one mask region utilizing a 4-connected seed fill operation.
20. A system for extracting highlighted regions in a scanned image document, comprising:
a data-processing apparatus comprising processor readable code embodied on one or more processor readable storage devices, said processor readable code comprising:
a module for converting a color space of a scanned image document into an sRGB color space;
a module for converting said sRGB color space into a plurality of background image planes and a plurality of mask binary regions;
a module for analyzing said plurality of background image planes utilizing at least one mask coordinate, wherein said plurality of background image planes are located beneath at least one mask binary region among said plurality of mask binary regions; and
a module for optically recognizing a highlighted text in said at least one mask binary region, by setting pixel values of uniform color pixels within said at least one mask binary region to white and remaining pixels within said at least one mask binary region to black if said plurality of background image planes beneath said at least one mask binary region comprises a uniform color.