Method and apparatus for compressing and decompressing images of documents

• US 5,432,870 A
• Filed: 06/30/1993
• Issued: 07/11/1995
• Est. Priority Date: 06/30/1993
• Status: Expired due to Term

• Current Assignee: Ricoh Company Limited, Ricoh Corporation (Ricoh Company Limited)
• Original Assignee: Ricoh Company Limited, Ricoh Corporation (Ricoh Company Limited)
• Inventors: Schwartz, Edward L.
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Tran, Phuoc
• Application Number:
  US08/086,184
• Time in Patent Office: 741 Days
• Field of Search: 382/56, 382/54, 382/50, 358/432, 358/433, 358/429, 358/262.1
• US Class Current: 382/232
• CPC Class Codes: H04N 1/41 Bandwidth or redundancy reduction by scanning H04N1/17 ; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00
• CPC Class Codes: H04N 19/30 using hierarchical techniques, e.g. scalability H04N19/63 takes precedence

Abstract

A method and apparatus for compressing images having a binary extractor, a binary compressor and a JPEG compressor. The binary extractor produces a binary image representing an estimate of the binary information in the original image. The binary image is compressed by the binary compressor. The binary image is also subtracted from the original image to produce a difference image representing the continuous-tone information in the original image. The difference image is then compressed using the JPEG compressor. In this manner, the present invention compresses a document image without having to perform segmentation on the image.

Claims

1. An image compression system for compressing an input image from an image source comprising:

a binary image generator to generate a binary image representing an estimate of the bi-level information in the input image, wherein the binary image includes edge information in the input image;

a post binary image generator coupled to the binary image generator to filter the binary image to produce a processed binary image;

a subtractor coupled to the post binary image generator to subtract the processed binary image from the input image to produce a difference image, wherein the processed binary image is subtracted from the input image on a pixel-by-pixel basis;

a post subtraction processing mechanism coupled to the subtractor to reduce components in the difference image created by mismatching between binary edges in the processed binary image and the input image to produce a resulting image, wherein portions of edges produced in the difference image due to fractional pixel misalignment between the processed binary image and the input image are removed from the processed binary image to create the resulting image;

a binary compressor coupled to the binary image generator to receive the binary image and compress the binary image into compressed binary data; and

continuous-tone compressor coupled to the post subtraction processing mechanism to compress the resulting image into compressed continuous-tone data, such that the input image is compressed into compressed binary data and compressed continuous-tone data.

2. The image compression system defined in claim 1 wherein the post binary image generator comprises a point spread modeling mechanism that convolves the binary image with an approximation of the point spread function of the image source to produce the processed image.

3. The image compression system defined in claim 1 wherein the post subtraction processing mechanism comprises a visual masking filter.

4. The image compression system defined in claim 1 wherein the post subtraction processing mechanism includes means for reducing high frequency edge mismatch errors.

5. The compression system defined in claim 1 wherein the continuous-tone image compressor comprises a Joint Photographic Experts Group (JPEG) compressor.

6. The compression system defined in claim 1 wherein the binary compressor comprises a Joint Binary Image Group (JBIG) compressor.

7. The compression system defined in claim 1 wherein the binary compressor comprises a Modified-Modified-READ (MMR) compressor.

8. The compression system defined in claim 1 wherein the binary compressor comprises an LZW compressor.

9. A compression system for compressing an input image comprising:

a binary image generator to generate a binary image representing an estimate of the bi-level information in the input image, wherein the binary image includes edge information in the input image;

a point spread modeling mechanism coupled to the binary image generator to receive the binary image and smooth the binary image to create a smoothed binary image;

a binary compressor coupled to the binary image generator to compress the binary image into compressed binary data;

a differencing mechanism coupled to receive the input image and the smoothed binary image to create a difference image, wherein the differencing mechanism includes a subtractor to subtract the smoothed binary image from the input image on a pixel-by-pixel basis;

a masking mechanism coupled to the differencing mechanism to reduce error in the difference image due to any fractional pixel misalignment of edges between the input image and the smoothed binary image to produce a masked image, wherein portions of edges produced in the difference image created by mismatching between binary edges in the input image and the smoothed binary image are removed to create the masked image; and

a continuous-tone compressor coupled to the masked image to compress the masked image into compressed continuous-tone data, such that the input image is compressed into compressed binary data and compressed continuous-tone data.

10. The compression system defined in claim 9 wherein the point spread modeling mechanism comprises a point spread filter.

11. The compression system defined in claim 9 wherein the binary image compressor includes means for determining blocks of the plurality of blocks that include edge information and means for binaizing all of the blocks in the input image.

12. The compression system defined in claim 11 wherein the means for determining the edge information in the plurality of blocks includes means for determining which of the plurality of blocks contain edge information and means for locating a representative binary level for the entire block.

13. The compression system as defined in claim 11 wherein the means for determining includes means for determining if a block contains a black pixel and a white pixel, such that each of the plurality of blocks is considered to have edge information if said each of the plurality of blocks contain a black pixel and a white pixel.

14. The compression system as defined in claim 9 wherein point spread modeling mechanism performs an area convolution on the binary image using an approximation of the point spread function of the scanner.

15. The compression system as defined in claim 14 wherein the area convolution is not performed when two adjacent blocks do not contain edge information.

16. The compression system as defined in claim 9 wherein the masking mechanism includes means for reducing the high frequency components in the difference image at edges in the binary image.

17. The compression system as defined in claim 9 wherein the masking mechanism comprises visual masking filter to filter the difference image.

18. The compression system defined in claim 9 wherein the continuous-tone image compressor comprises a Joint Photographic Experts Group (JPEG) compressor.

19. The compression system defined in claim 9 wherein the binary compressor comprises a Joint Binary Image Group (JBIG) compressor.

20. The compression system defined in claim 9 wherein the binary compressor comprises a Modified-Modified-READ (MMR) compressor.

21. The compression system defined in claim 9 wherein the binary compressor comprises a LZW compressor.

22. A compression system for compressing an input image comprising:

a binary image generator for generating a binary image representing an estimate of the bi-level information in the input image, wherein the binary image includes edge information in the input image;

a point spread model coupled to the binary image generator for receiving the binary image and smoothing the binary image to create a smoothed binary image;

a binary compressor coupled to the binary image generator for compressing the binary image into compressed binary data;

a subtracter coupled to receive the input image and the smoothed binary image to create a difference image on a pixel-by-pixel basis;

visual masking filter coupled to the subtracter for reducing error in the difference image due to any fractional pixel misalignment of edges between the input image and the smoothed binary image; and

continuous-tone compressor coupled to the visual making filter for compressing the masked image, such that mosquito noise is reduced when the input image is reconstructed.

23. A decompression system for decompressing image data that included compressed binary image data and compressed continuous-tone image data into a reconstructed version of the original image, said system comprising:

a binary decompressor to decompress the compressed binary image data to produce a binary image;

a modeling mechanism coupled to the binary decompressor to receive the binary image and process the binary image to create a processed binary image;

a continuous-tone image decompressor to decompress the compressed continuous-tone image data to produce a continuous-tone image; and

a combining mechanism to combine the processed binary image and the continuous-tone image on a pixel-by-pixel basis to produce the reconstructed version of the original image.

24. The system defined in claim 23 wherein the continuous-tone decompressor comprises a JPEG decompressor.

25. The system defined in claim 23 wherein the modeling mechanism includes a point spread modeling mechanism that performs an area convolution on the binary image using an approximation of the point spread function of the source of the original image.

26. The compression system as defined in claim 25 wherein the area convolution is not performed when two adjacent blocks do not contain edge information.

27. A decompression system for decompressing image data corresponding to an original image, wherein the image data includes compressed binary image data and compressed continuous-tone image data, and wherein the image data is decompressed into a reconstructed version of the original image, said system comprising:

a binary decompressor coupled to receive the compressed binary image data for decompressing the compressed binary image data to produce a binary image having a first plurality of pixels;

a point spread filter coupled to receive the binary image for performing an area convolution on the binary image using an approximation of the point spread function of the source of the original image to create a smoothed binary image;

a continuous-tone image decompressor coupled to receive the continuous-tone image data wherein the continuous-tone decompressor decompresses the compressed continuous-tone image data to produce a continuous-tone image having a second plurality of pixels; and

an adder coupled to receive the smoothed binary image and the continuous-tone image for combining the smoothed binary image and the continuous-tone image on a pixel-by-pixel basis to produce the reconstructed version of the original image.

28. A reproduction system for reproducing an image source comprising:

scanning means for scanning the image source into image data;

compression means coupled to the scanning means for compressing the image data, wherein the compression means includes

binary image generation means for generating a binary image representing an estimate of the bi-level information in the input image, wherein the binary image includes edge information in the input image;

post binary image generation processing means coupled to the binary image generation means for filtering the binary image to produce a smoothed binary image;

subtraction means coupled to the post binary image generation processing means for subtracting the processed binary image from the input image to produce a difference image wherein the processed binary image is subtracted from the input image on a pixel-by-pixel basis;

post subtraction processing means coupled to the subtraction means for reducing components in the difference image created by mismatching between binary edges in the processed binary image and the input image to produce a resulting image., wherein portions of edges produced in the difference image due to fractional pixel misalignment between the processed binary image and the input image are removed from the processed binary image to create the resulting image;

binary compression means coupled to the binary image generation means for receiving the binary image and compressing the binary image into compressed binary data; and

continuous-tone compression means coupled to the post subtraction processing means for compressing the resulting image into compressed continuous-tone data, such that the input image is compressed into compressed binary data and compressed continuous-tone data,

decompression means coupled to the binary compression means and continuous-tone compression means for decompressing the compressed binary image data and the compressed continuous-tone image data into reconstructed image data;

display means for producing a visual display of the image data.

29. The reproduction system defined in claim 28 wherein the display means comprises a printing means for producing a hard copy of the image data.

30. A document archiving system for use with a document image comprising:

input means for producing image data representing the document image;

compression means coupled to the input means for compressing the image data, wherein the compression means includes

binary image generation means for generating a binary image representing an estimate of the bi-level information in the input image, wherein the binary image includes edge information in the input image,

post binary image generation processing means coupled to the binary image generation means for filtering the binary image to produce a smoothed binary image,

subtraction means coupled to the post binary image generation processing means for subtracting the smoothed binary image from the input image to produce a difference image wherein the smoothed binary image is subtracted from the input image on a pixel-by-pixel basis;

post subtraction processing means coupled to the subtraction means for reducing high frequency components in the difference image created by mismatching between binary edges in the smoothed binary image and the input image to produce a resulting image, wherein portions of edges produced in the difference image due to fractional pixel misalignment between the smoothed binary image and the input image am removed from the processed binary image to create the resulting image;

binary compression means coupled to the binary image generation means for receiving the binary image and compressing the binary image into compressed binary image data into compressed binary data; and

continuous-tone compression means coupled to the post subtraction processing means for compressing the resulting image into compressed continuous-tone image data; and

storage means for storing the document image as compressed binary image data and compressed continuous-tone image data.

31. The system defined in claim 30 further comprising a retrieval means for retrieving the document image from the storage means, wherein the retrieval means includes

binary decompression means for decompressing the compressed binary image data to produce a binary image;

modeling means coupled to the binary decompression means for receiving the binary image and smoothing the binary image to create a smoothed binary image;

continuous-tone image decompression means for decompressing the compressed continuous-tone image data to produce a continuous-tone image; and

combining means for combining the smoothed binary image and the continuous-tone image to produce a reconstructed version of the original document image.

32. A facsimile system for transferring a document image comprising:

scanning means for inputting the document image as image data;

transmission means for transmitting the image data, wherein the transmission means includes compression means coupled to the scanning means for compressing the image data and channel means for sending the image data, wherein the compression means includes

binary image generation means for generating a binary image representing an estimate of the bi-level information in the input image, wherein the binary image includes edge information in the input image,

post binary image generation processing means coupled to the binary image generation means for filtering the binary image to produce a smoothed binary image,

subtraction means coupled to the post binary image generation processing means for subtracting the smoothed binary image from the input image to produce a difference image, wherein the smoothed binary image is subtracted from the input image on a pixel-by-pixel basis;

post subtraction processing means coupled to the subtraction means for reducing high frequency components in the difference image created by mismatching between binary edges in the smoothed binary image and the input image to produce a resulting image, wherein portions of edges produced in the difference image due to fractional pixel misalignment between the smoothed binary image and the input image are removed from the processed binary image to create the resulting image;

binary compression means coupled to the binary image generation means for receiving the binary image and compressing the binary image into compressed binary image data into compressed binary data; and

continuous-tone compression means coupled to the post subtraction processing means for compressing the resulting image into compressed continuous-tone image data, such that the channel means sends the compressed binary image data and compressed continuous-tone image data.

33. The system defined in claim 32 further comprising reception means for receiving the document image, wherein the reception means includes:

input means for receiving the compressed binary image data and the compressed continuous-tone image data;

binary decompression means coupled to the input means for decompressing the compressed binary image data to produce a binary image,

modeling means coupled to the binary decompression means for receiving the binary image and smoothing the binary image to create a smoothed binary image,

continuous-tone image decompression means coupled to the input means for decompressing the compressed continuous-tone image data to produce a continuous-tone image,

combining means for combining the smoothed binary image and the continuous-tone image on a pixel-by-pixel basis to produce a reconstructed version of the original document image; and

printing means coupled to the combining means for producing a hard copy representing the document image.