Method for adaptively masking off a video window in an overscanned image

• US 4,774,569 A
• Filed: 07/24/1987
• Issued: 09/27/1988
• Est. Priority Date: 07/24/1987
• Status: Expired due to Term

• Current Assignee: Eastman Kodak Company
• Original Assignee: Eastman Kodak Company
• Inventors: Morton, Roger R. A., Redden, John E.
• Primary Examiner(s): Britton, Howard W.
• Application Number:
  US07/077,718
• Time in Patent Office: 431 Days
• Field of Search: 358/282, 358/280, 358/102, 382/22
• US Class Current: 348/107
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: H04N 1/38 Circuits or arrangements for blanking or otherwise eliminating unwanted parts of pictures H04N1/387 takes precedence

Abstract

A method for extracting (masking off) a desired video window, such as that which forms a scanned image of a document or a portion thereof, from a overscanned image stored in a frame buffer is disclosed herein. Specifically, this method involves first determining horizontal pixel locations for left and right edges of the document for each of a plurality of scan lines located within said scanned image. Then, histograms of the left and right edge locations are fabricated. Once this has occurred, the positions of top and bottom edges of an escribing rectangle is found. The escribing rectangle is a rectangular area, located within the frame buffer, which includes the entire document. Now, with the top and bottom edge locations known, the inventive method locates the right and left edge locations of the escribing rectangle by suitably analyzing the histogram results. Next, memory address corresponding to the first pixel location in the escribing rectangle and the vertical and horizontal size of the rectangle are determined, using all the edge locations of the escribing rectangle, and then applied as input to addressing circuitry located within the frame buffer. Once this occurs, then all (or a portion of) the pixel values stored within the frame buffer and located on and within the escribing rectangle are successively accessed and routed to an output connection as being those pixel values that form the image, or a selected portion thereof, of the document.

Claims

1. In a system for processing a scanned image stored within a frame buffer, wherein the scanned image contains pixels that collectively form an image of a document, a method for extracting pixel values from said frame buffer and associated with at least a portion of the image of the document, said method comprising the steps of:

determining horizontal pixel locations for left and right edges of the image of the document for each of a plurality of scan lines located within said scanned image, wherein each of said scan lines contains at least one pixel that comprises the image of the document;

ascertaining, in response to the left and right edges in said plurality of scan lines, first and second vertical pixel locations of corresponding top and bottom edges of an escribing rectangle located within a frame buffer that stores said scanned image, wherein said escribing rectangle completely contains said image of the document;

determining, in response to said horizontal pixel locations for said left edge, a location where a first pre-determined percentage of said left edges has occurred in order to select a first horizontal pixel location in said scanned image at which a leading edge of the escribing rectangle occurs;

determining, in response to said horizontal pixel locations for said right edge, a location where a second pre-determined percentage of said right edges has occurred in order to select a second horizontal pixel location in said scanned image at which a trailing edge of said escribing rectangle occurs;

determining, in response to said first and second horizontal pixel locations and said first and second vertical pixel locations, memory addressing information relating to where pixel values, that are contained within said escribing rectangle, are located within said frame buffer;

accessing, in response to said information, contents of various memory locations, stored within said frame buffer, in order to obtain selected pixel values; and

providing said selected pixel values to an output connection as the pixel values that are associated with at least a portion of the image of the document.

2. The method in claim 1 wherein said memory addressing information determining step comprises the steps of:

determining a memory location of a first pixel in the escribing rectangle and horizontal and vertical size information of said rectangle; and

applying said first pixel memory location and the vertical and horizontal size information to addressing circuitry situated within said frame buffer.

3. The method in claim 2 wherein said left and right edge pixel locations determining step comprises the step of:

comparing a separation between the left and right edges of a current scan line within said plurality of scan lines to a pre-determined amount and, if said separation is less than said pre-determined amount, selecting left and right edge locations for a most recent scan line in said plurality of scan lines for the left and right edge locations for the current scan line.

4. The method in claim 3 wherein said first and second vertical pixel locations ascertaining step includes the step of finding the top edge of said escribing rectangle as being a vertical pixel address of a first scan line in a first group of a first pre-determined number of successive scan lines, wherein each of the said scan lines in said first group contains a leading edge.

5. The method in claim 4 wherein said first and second vertical pixel locations ascertaining step includes the step of finding the bottom edge of said escribing rectangle as being a vertical pixel address of a first scan line in a second group of a second pre-determined number of successive scan lines, wherein each of said scan lines in said second group does not contain a trailing edge.

6. The method in claim 5 wherein said first and second pre-determined percentages are both approximately equal to 10%.

7. The method in claim 6 wherein said scanned image is a scanned microfilm image.

8. The method in claim 1 wherein said left and right edge pixel locations determining step comprises the steps of:

providing, in response to at least one incoming pixel value that comprises a scan line in the scanned image, a background value representative of a background level in said scanned image and associated with said incoming pixel value;

generating, in response to said background value and to said incoming pixel value, a first signal indicating that said background value exceeds a first value and a second signal indicating that said incoming pixel value exceeds a second value;

determining, in response to said first and second signals, whether an edge transition has occurred in the scan line and, in response thereto, providing a suitable indication thereof, wherein said determining and providing step comprises the steps of;

ascertaining, as a pre-defined logical function of said first and second signals, whether said edge transition has occurred and, in response thereto, producing at least one suitable control signal; and

determining, in response to said one control signal and to a pixel clock signal, whether said edge transition is followed in said scan line by a pre-defined number of similar pixel values and, in response thereto, providing said indication;

storing, in response to said indication and to a horizontal pixel address representing a pixel location of said incoming pixel on said scan line, said pixel address as a stored edge; and

selecting pixel locations for a leading edge and a corresponding trailing edge for said scan line, from amongst the stored edges, wherein said selected pixel locations for the leading and trailing edges have a widest pixel spacing of all the leading and corresponding trailing edges stored for said scan line in said scanned image.

9. The method in claim 8 wherein said first and second signal generating step comprises the steps of:

generating first and second comparison levels, wherein said first and second comparison levels slowly change in order to track variations in said background level and said incoming pixel values;

comparing said background value to the first comparison level and comparing said incoming pixel values to said second comparison level in order to generate said first and second signals, respectively.

10. The method in claim 9 wherein said edge transition ascertaining step comprises the step of ignoring any edge transition that is not followed by at least a pre-determined minimum number of similar pixel values in order to filter noise from said image.

11. The method in claim 10 wherein said pre-determined minimum number is "5" and said pre-defined number is "20".

12. The method in claim 11 wherein said memory addressing information determining step comprises the steps of:

determining a memory location of a first pixel in the escribing rectangle and horizontal and vertical size information of said rectangle; and

applying said first pixel memory location and the vertical and horizontal size information to addressing circuitry situated within said frame buffer.

13. The method in claim 12 wherein said left and right edge pixel location determining step comprises the step of:

comparing a separation between the left and right edges of a current scan line within said plurality of scan lines to a pre-determined amount and, if said separation is less than said pre-determined amount, selecting left and right edge locations for a most recent scan line in said plurality of scan lines for the left and right edge locations for the current scan line.

14. The method in claim 13 wherein said first and second vertical pixel locations ascertaining step includes the step of finding the top edge of said escribing rectangle as being a vertical pixel address of a first scan line in a first group of a first pre-determined number of successive scan lines, wherein each of the said scan lines in said first group contains a leading edge.

15. The method in claim 14 wherein said first and second vertical pixel locations ascertaining step includes the step of finding the bottom edge of said escribing rectangle as being a vertical pixel address of a first scan line in a second group of a second pre-determined number of successive scan lines, wherein each of said scan lines in said second group does not contain a trailing edge.

16. The method in claim 15 wherein said first and second pre-determined percentages are both approximately equal to 10%.

17. The method in claim 16 wherein said scanned image is a scanned microfilm image.