Local adaptive contrast enhancement

• US 5,524,070 A
• Filed: 08/15/1994
• Issued: 06/04/1996
• Est. Priority Date: 10/07/1992
• Status: Expired due to Fees

• Current Assignee: The Research Foundation for The State University of New York (State University of New York)
• Original Assignee: The Research Foundation for The State University of New York (State University of New York)
• Inventors: Shin, Yong-Chul, Srihari, Sargur N., Demjanenko, Victor, Sridhar, Ramalingam
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): ANDERSON, DAVID
• Application Number:
  US08/290,173
• Time in Patent Office: 659 Days
• Field of Search: 382/27, 382/18, 382/54, 382/51, 382/50, 382/52, 382/41, 382/254, 382/266, 382/261, 382/263, 382/274, 382/272
• US Class Current: 382/274
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/162 Quantising the image signal

Abstract

In a mail sorting system 10, output pixel intensities of an optical scanner 18 have their contrast locally enhanced. A contrast enhancer 24 uses statistical methods (averaging, standard deviation) coupled with empirical stretch and offset data stored in a ROM 25, to enhance pixel contrast in a pipelined processing operation.

Claims

1. A method for enhancing contrast of pixels of an image having foreground and background portions comprising the steps of,

scanning the image;

dividing the scanned image into a plurality of pixels;

generating a series of data signals, each signal having an intensity magnitude proportional to the amount of foreground and background in each pixel;

selecting a first number of pixel data signals defining a first window of pixels and a corresponding center pixel in said window;

storing in a memory values for offset and stretch corresponding to predetermined ranges of standard deviation and average intensity;

computing the average intensity and standard deviation of the selected pixel data signals in the window;

accessing said memory to retrieve a stored value of stretch corresponding to a computed standard deviation in a range between zero and approximately zero and corresponding to a computed average intensity in a range much greater than zero;

accessing said memory to retrieve a stored value of offset corresponding to a computed standard deviation in a range between zero and approximately zero and corresponding to a computed average intensity in a range between zero and approximately zero; and

selectively changing the intensity magnitude of a pixel data signal for a pixel disposed near the center pixel of said window in accordance with accessed values of offset and stretch.

2. The method of claim 1 further comprising selecting a second number of pixels data signals defining a second window of pixels with at least one pixel of said second window being different from the pixel of the first window;

generating another series of data signals, each signal having an intensity magnitude proportional to the amount of foreground and background in each pixel of said second window;

computing the average intensity of said first window and computing the standard deviation of said second window;

and selectively changing the intensity magnitude of the data signal of the center pixel of the second window.

3. The method of claim 2 wherein both windows have the same center pixel.

4. The method of claim 1 where the changed pixel data signal is the data signal of the center pixel.

5. The method of claim 1 further comprising the step of serially moving the window a distance of one pixel to respectively define new center pixels and new windows and enhancing the contrast of each new center pixel.

6. The method of claim 1 further comprising the step of selectively changing the intensity magnitude of the data signal of the center pixel when the average intensity of the window is in a first range between zero and approximately zero or in a second range much greater than zero,

7. The method of claim 1 further comprising the step of selectively changing the intensity magnitude of the data signal of the center pixel when the standard deviation is in a range between zero and approximately zero.

8. The method of claim 1 wherein the intensity magnitude of the data signal of the center pixel is stretched when the standard deviation is in a range between zero and approximately zero and the average intensity is in a range of much greater than zero.

9. The method of claim 1 wherein the intensity magnitude of the data signal of the center pixel is offset when the standard deviation and average intensity are each in a range between zero and approximately zero.

10. The method of claim 8 wherein the stretch values comprise predetermined amounts for selected ranges of standard deviation and average intensity.

11. The method of claim 9 wherein the offset values comprise predetermined amounts for selected ranges of standard deviation and average intensity.

12. The method of claim 1 wherein the ranges of average intensity overlap.

13. A system for enhancing contrast of pixels of an image having foreground and background portions comprising,

an optical image scanner for scanning an image and generating a series of pixel data signals having magnitudes representative of the intensity of each pixel in the image,

means the selecting a first number of pixel data signals to define a first window of pixel signals and a corresponding center pixel signal in said window,

means for computing the average intensity and standard deviation of the pixel signals of said window,

a read-only memory for storing values of offset and stretch corresponding to a predetermined range of standard deviation between zero and approximately zero and predetermined ranges of average intensity, including a first average intensity range between zero and approximately zero and a second average intensity range of much greater than zero, and

means for enhancing contrast by selectively changing the intensity of a pixel data signal of a pixel near the center pixel in accordance with the accessed values of offset and stretch.

14. The method of claim 13 further comprising selecting a second number of pixels data signals defining a second window of pixels with at least one pixel of said second window being different from the pixels of said first window;

computing the average intensity of the magnitude of said first window signals and computing the standard deviation of said second window signals;

and selectively changing the intensity of the center pixel data signals of the second window.

15. The method of claim 14 wherein both windows have the same center pixel.

16. The system of claim 13 wherein the changed pixel data signal is the data signal of the center pixel.

17. The system of claim 13 further comprising means for moving the pixel window a distance of one pixel to serially define subsequent new windows.

18. The system of claim 13 wherein said means for selectively changing the center pixel data signal intensity magnitude changes said magnitude when the standard deviation is zero or approximately zero.

19. The system of claim 13 wherein said means for selectively changing the pixel data signal intensity magnitude changes said magnitude when the average intensity is zero or approximately zero or much greater than zero.

20. The system of claim 13 wherein said means for changing the pixel data signal intensity magnitude stretches intensity of the pixel data signal when the standard deviation of the window is zero or approximately zero and the average intensity of the window is much greater than zero.

21. The system of claim 13 wherein said means for changing the pixel data signal intensity magnitude offsets the magnitude of the pixel data signal when the standard deviation is zero or approximately zero and the average intensity is zero or approximately zero.

22. The system of claim 20 wherein the stored stretch values are a predetermined amount for selected ranges of standard deviation and average intensity.

23. The system of claim 21 wherein the stored offset values are predetermined amounts for selected ranges of standard deviation and average intensity.

24. The method of claim 13 wherein the ranges of average intensity overlap.

25. In a system for contrast enhancement of pixel data, a chip set, including a first integrated circuit comprising,

an input for receiving data signals representative of pixel intensity,

a first circuit for squaring the value of pixel intensity data signals,

a second circuit for summing the squared values of pixel intensity data signals,

a third circuit for summing the input pixel intensity data signals,

a fourth circuit for squaring the sum of input pixel intensity data signals,

a plurality of shift registers connected to said second, third and fourth circuit means for receiving data representative of the sum of the input pixel intensity data, the sum of the squares of input pixel intensity data, and the square of the sum of input pixel intensity data,

an averaging and standard deviation calculating circuit coupled to the shift registers for computing the average and standard deviation of input pixel intensity data,

a second integrated circuit comprising a memory having stored values of offset and stretch, said stored values of offset and stretch being stored in accordance with a range of standard deviation and average intensity, and

a third integrated circuit comprising a pixel intensity data alteration circuit coupled to the first and second integrated circuits, for altering the magnitude of the input pixel intensity data signals in accordance with the standard deviation and the average intensity to generate output signals representative of locally contrast enhanced pixel intensity data.

26. The chip set of claim 25 wherein the first and second ranges of average intensity stored in the second circuit overlap.

27. A method for enhancing contrast of pixels of an image having foreground and background portions comprising the steps of,

scanning the image;

dividing the scanned image into a plurality of pixels;

generating a series of data signals, each signal having an intensity magnitude proportional to the amount of foreground and background in each pixel;

selecting a first number of pixel data signals defining a first window of pixels and a corresponding center pixel in said window;

storing in a memory values for offset and stretch corresponding to predetermined ranges of standard deviation and average intensity;

computing the average intensity and standard deviation of the selected pixel data signals in the window;

accessing said memory when the computed standard deviation is in a range between zero and approximately zero;

retrieving from said memory a stored offset value when said computed average intensity is in a range between zero and approximately zero;

retrieving from said memory a stored stretch value when said computed average intensity is in a range much greater than zero; and

selectively changing the intensity magnitude of a pixel data signal for a pixel disposed near the center pixel of said window in accordance with retrieved value.