Image coding method and apparatus for localized decoding at multiple resolutions

• US 7,088,866 B2
• Filed: 01/26/2004
• Issued: 08/08/2006
• Est. Priority Date: 07/03/1998
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Andrew, James Philip
• Primary Examiner(s): Do, Anh Hong
• Application Number:
  US10/763,213
• Publication Number:
  US 20040170332A1
• Time in Patent Office: 925 Days
• Field of Search: 382/233, 382/232, 382/244, 382/245, 382/246, 382/248, 382/250, 382/40, 386/95, 375/240.15
• US Class Current: 382/233
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/162 User input
• CPC Class Codes: H04N 19/17 the unit being an image region, e.g. an object
• CPC Class Codes: H04N 19/36 Scalability techniques involving formatting the layers as a function of picture distortion after decoding, e.g. signal-to-noise [SNR] scalability
• CPC Class Codes: H04N 19/63 using sub-band based transform, e.g. wavelets
• CPC Class Codes: H04N 19/645 by grouping of coefficients into blocks after the transform

Abstract

The method generates a compressed digital image from a original image. The compressed digital image allows random access to portions of the compressed image at a number of resolutions. The original image is first transformed (103) by a multi-level DWT to form a non redundant multiple resolution frequency domain representation of the image. The representation comprises a DC subband and a plurality of high frequency subbands arranged as levels, where each level represents a frequency contribution between adjacent resolutions and where each subband comprises a plurality of tiles. The DC subband is then entropy encoded (104) into the bitstream. The high frequency subbands are next entropy encoded (108,109, and 110) into the bitstream in level order (105,113,114) and tile order (107,111).

Claims

1. A method of decoding a sequential stream of data for providing substantially random access to portions of an image at a plurality of predetermined resolutions, the stream including a non-redundant multiple resolution frequency domain representation of the image, the representation comprising one low frequency subband and a plurality of high frequency subbands arranged in levels, wherein each subband is divided into a plurality of tiles and each level represents frequency contributions between adjacent resolutions of the image, and wherein each tile represents a frequency contribution to a portion of the image at a predetermined resolution, said method including:

retrieving at most one pointer in the sequential stream for each set of tiles which correspond to substantially a same spatial portion of the image for each level to access a portion of the digital image.

2. The method as claimed in claim 1, wherein accessing the portion includes decoding a data address by the pointer.

3. The method as claimed in claim 1, wherein the sequential stream is a bitstream.

4. An apparatus for decoding a sequential stream of data for providing substantially random access to portions of an image at a plurality of predetermined resolutions, the stream including a non-redundant multiple resolution frequency domain representation of the image, the representation comprising one low frequency subband and a plurality of high frequency subbands arranged in levels, wherein each subband is divided into a plurality of tiles and each level represents frequency contributions between adjacent resolutions of the image, and wherein each tile represents a frequency contribution to a portion of the image at a predetermined resolution, said apparatus including:

means for retrieving at most one pointer in the sequential stream for each set of tiles which correspond to substantially a same spatial portion of the image for each level to access a portion of the digital image.

5. The apparatus as claimed in claim 4, wherein accessing the portion includes decoding a data address by the pointer.

6. The apparatus as claimed in claim 4, wherein the sequential stream is a bitstream.

7. A computer program product comprising a computer readable medium having recorded thereon a computer program for decoding a sequential stream of data for providing substantially random access to portions of an image at a plurality of predetermined resolutions, the stream including a non-redundant multiple resolution frequency domain representation of the image, the representation comprising one low frequency subband and a plurality of high frequency subbands arranged in levels, wherein each subband is divided into a plurality of tiles and each level represents frequency contributions between adjacent resolutions of the image, and wherein each tile represents a frequency contribution to a portion of the image at a predetermined resolution, said computer program product including:

code for a retrieval step, of retrieving at most one pointer in the sequential stream for each set of tiles which correspond to substantially a same spatial portion of the image for each level to access a portion of the digital image.

8. The computer program product claimed in claim 7, wherein accessing the portion includes decoding a data address by the pointer.

9. The computer program product as claimed in claim 7, wherein the sequential stream is a bitstream.

10. A method of decoding a bitstream, said bitstream comprising a non-redundant hierarchical code, having one low frequency subband and a plurality of high frequency subbands arranged in levels, which levels combine to represent multiple resolutions of the image, and each subband is divided into a plurality of tiles which are entropy encoded, said method comprising the steps of:

retrieving from the bitstream those entropy encoded tiles representing substantially the same predetermined spatial portion of the image from a number of contiguous levels of entropy encoded tiles representing a predetermined resolution of the multiple resolutions of the digital image, wherein at least two entropy coded tiles from different the subbands which represent substantially the same predetermined spatial portion of the image are arranged contiguously in the bitstream;

entropy decoding each tile;

grouping the tiles into a plurality of subbands; and

inverse linear transforming the grouped subbands to produce the predetermined portion of the digital image at the predetermined resolution.

11. An apparatus for decoding a bitstream, the bitstream comprising a non-redundant hierarchical code, having one low frequency subband and a plurality of high frequency subbands arranged in levels, which levels combine to represent multiple resolutions of the image, and each subband is divided into a plurality of tiles which are entropy encoded, said apparatus comprising:

means for retrieving from the bitstream those entropy encoded tiles representing substantially the same predetermined portion of the image from a number of contiguous levels of entropy encoded tiles representing a predetermined resolution of the multiple resolutions of the digital image, wherein at least two entropy coded tiles from different subbands which represent substantially the same predetermined spatial portion of the image are arranged contiguously in the bitstream;

means for entropy decoding each tile;

means for grouping the tiles into a plurality of subbands; and

means for inverse linear transforming the grouped subbands to produce the predetermined portion of the digital image at the predetermined resolution.

12. A computer readable medium comprising a computer program for decoding a bitstream, the bitstream comprising a non-redundant hierarchical code, having one low frequency subband and a plurality of high frequency subbands arranged in levels, which levels combine to represent multiple resolutions of the image, and each subband is divided into a plurality of tiles which are entropy encoded, said computer program comprising:

code for a retrieval step, of retrieving from the bitstream those entropy encoded tiles representing substantially the same predetermined portion of the image from a number of contiguous levels of entropy encoded tiles representing a predetermined resolution of the multiple resolutions of the digital image, wherein at least two entropy coded tiles from different subbands which represent substantially the same predetermined spatial portion of the image are arranged contiguously in the bitstream;

code for an entropy decoding step, of entropy decoding each tile;

code for a grouping step, of grouping the tiles into a plurality of subbands; and

code for a transforming step, of inverse linear transforming the grouped subbands to produce the predetermined portion of the digital image at the predetermined resolution.

13. A method of decoding a portion of an image from a bitstream, the bitstream being generated from a plurality of frequency subbands created by wavelet transforming a digital image, each of the subbands being divided into a plurality of tiles, each of the tiles being entropy encoded and arranged so that the entropy-encoded tiles representing substantially a same part of the image are arranged contiguously in the bitstream for at least one wavelet transform level, said method comprising the steps of:

selecting the portion of the image at a selected resolution;

accessing the entropy-encoded tiles of the bitstream corresponding to the selected resolution and selected portion of the image; and

decoding the image portion using the accessed entropy-encoded tiles.

14. An apparatus for decoding a portion of an image from a bitstream, the bitstream being generated from a plurality of frequency subbands created by wavelet transforming a digital image, each of the subbands being divided into a plurality of tiles, each of the tiles being entropy encoded and arranged so that the entropy-encoded tiles representing substantially a same part of the image are arranged contiguously in the bitstream for at least one wavelet transform level, said apparatus comprising:

means for selecting the portion of the image at a selected resolution;

means for accessing the entropy-encoded tiles of the bitstream corresponding to the selected resolution and selected portion of the image; and

means for decoding the image portion using the accessed entropy-encoded tiles.

15. A computer program product comprising machine-readable program code recorded on a machine-readable recording medium, for controlling the operation of a data processing apparatus on which the program code executes to perform a method of decoding a portion of an image from a bitstream, the bitstream being generated from a plurality of frequency subbands created by wavelet transforming a digital image, each of the subbands being divided into a plurality of tiles, each of the tiles being entropy encoded and arranged so that the entropy-encoded tiles representing substantially a same part of the image are arranged contiguously in the bitstream for at least one wavelet transform level, said method comprising the steps of:

selecting the portion of the image at a selected resolution;

accessing the entropy-encoded tiles of the bitstream corresponding to the selected resolution and selected portion of the image; and

decoding the image portion using the accessed entropy-encoded tiles.

16. A system for decoding a portion of an image from a bitstream, said system comprising:

a storage unit for storing the bitstream, wherein the bitstream is generated from a plurality of frequency subbands created by wavelet transforming a digital image, each of the subbands being divided into a plurality of tiles, each of the tiles being entropy encoded and arranged so that the entropy-encoded tiles representing substantially a same part of the image are arranged contiguously in the bitstream for at least one wavelet transform level; and

a processor in communication with said storage unit and adapted to perform a method of decoding the portion of the image from the bitstream, the method comprising the steps of;

receiving a selection of the portion of the image and a selected resolution;

accessing the entropy-encoded tiles of the bitstream corresponding to the selected resolution and selected portion of the image; and

decoding the image portion using the accessed entropy-encoded tiles.