Image processing using probabilistic local behavior assumptions

• US 20050084166A1
• Filed: 06/25/2003
• Published: 04/21/2005
• Est. Priority Date: 06/25/2002
• Status: Active Grant

• Current Assignee: QUIP Technology Limited
• Original Assignee: QUIP Technology Limited
• Inventors: Boneh, Ran, Toib, Yotam
• Granted Patent: US 7,474,794 B2
• Time in Patent Office: Days
• US Class Current: 382/239
• CPC Class Codes: G06T 9/20 Contour coding, e.g. using detection of edges
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/134 characterised by the element, parameter or criterion affecting or controlling the adaptive coding
• CPC Class Codes: H04N 19/17 the unit being an image region, e.g. an object
• CPC Class Codes: H04N 19/172 the region being a picture, frame or field
• CPC Class Codes: H04N 19/1883 the unit relating to sub-band structure, e.g. hierarchical level, directional tree, e.g. low-high [LH], high-low [HL], high-high [HH]
• CPC Class Codes: H04N 19/34 Scalability techniques involving progressive bit-plane based encoding of the enhancement layer, e.g. fine granular scalability [FGS]
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/619 the transform being operated outside the prediction loop
• CPC Class Codes: H04N 19/63 using sub-band based transform, e.g. wavelets
• CPC Class Codes: H04N 19/635 characterised by filter definition or implementation details

Abstract

Image transform apparatus for processing pixel image data, comprising: at least one low pass extractor for extracting a low pass component from a series of image pixels, at least one low pass downsampler associated with said at least one low pass extractor for producing downsampled versions of respective extracted low pass components, at least one high pass extractor for extracting a high pass component from a series of image pixels, at least one high pass downsampler associated with said at least one high pass extractor for producing downsampled versions of respective extracted low pass components, and control and iteration circuitry associated with said low pass extractor and said high pass extractor for: feeding complementary sets of pixels from an image to said respective extractors, receiving respective low and high pass components from respective high and low pass downsamplers, and iteratively feeding complementary sets of pixels from a downsampled version of said most recently formed low pass component to respective high and low pass extractors and receiving further respective low and high pass components until a stop condition is reached. The corresponding reverse transform uses assumptions about localized step or edge behavior in the image.

Claims

1. Image transform apparatus for processing pixel image data, comprising:

at least one low pass extractor for extracting a low pass component from a series of image pixels,

at least one low pass downsampler associated with said at least one low pass extractor for producing downsampled versions of respective extracted low pass components,

at least one high pass extractor for extracting a high pass component from a series of image pixels,

at least one high pass downsampler associated with said at least one high pass extractor for producing downsampled versions of respective extracted low pass components, and

control and iteration circuitry associated with said low pass extractor and said high pass extractor for;

feeding complementary sets of pixels from an image to said respective extractors,

receiving respective low and high pass components from respective high and low pass downsamplers, and

iteratively feeding complementary sets of pixels from a downsampled version of a most recently formed low pass component to respective high and low pass extractors and receiving further respective low and high pass components until a stop condition is reached.

2. Apparatus according to claim 1, further comprising data storage circuitry for storing each high pass component and a final low pass component.

3. Apparatus according to claim 1, further comprising a stop condition input for setting said stop condition.

4. Apparatus according to claim 3, wherein said stop condition input is associated with a user interface.

5. Apparatus according to claim 3, wherein said stop condition input is operable to receive data regarding at least one of image importance, image part importance, and available bandwidth, and to use said data to set said stop condition dynamically.

6. Apparatus according to claim 1, further comprising video frame processing circuitry for providing said pixel image data from a video frame.

7. Apparatus according to claim 6, wherein said frame is one of a series of frames, the apparatus being operable to obtain said high and low pass components for each frame in said series and further comprising frame selection circuitry to select video frames for feeding to a compressor to compress said frames as a video stream, thereby to produce a video stream of a required frame rate.

8. Apparatus according to claim 1, wherein said pixel image data and said stop condition are set such that said processing comprises lossless image compression.

9. Apparatus according to claim 1, wherein said stop condition is variable such that said processing provides different levels of lossy compression to said pixel image data.

10. Image transform apparatus for recovering an original image from a series of successively obtained high pass components and a final low pass component, the apparatus comprising:

an estimator comprising;

a component combiner for combining a current level high pass component and a current low pass component to form a combination,

an upsample guess unit for estimating an upsampled version of said combination and iteratively correcting said version to converge on a next level image, said upsampler comprising an edge enhancement unit configured with a step function to scan said upsampled image, said step function being able to identify and enhance step like local behavior over said image, thereby to enhance resolution of said upsampled image, said image providing a next level low pass component, and

a control and iteration unit for iteratively feeding to said estimator a next one of said successively obtained high pass components and a respective next level low pass component until a zeroth level low pass component is reached.

11. Apparatus according to claim 10 wherein said upsample guess unit is configured to use said high pass components to identify said step-like local behavior, thereby to carry out said enhancing.

12. Apparatus according to claim 10, wherein said step function is a non-linear step function.

13. Apparatus according to claim 10, wherein said step function comprises an edge enhancement function.

14. Apparatus according to claim 10, wherein said control and iteration unit further comprises upsampling functionality for controlling said estimator to operate said edge unit on said zeroth level low pass component to provide an enhancement thereof.

15. Apparatus according to claim 14, wherein said upsampling functionality is usable on an original image to provide an enhanced resolution version thereof.

16. A method of transforming a pixilated image comprising:

levelwise iterative stages of;

extracting, downsampling and retaining high frequency information components, and

generating low frequency components for extraction therefrom of further high frequency components, and

downsampling said low frequency components, for a next level of said iteration,

until a stop condition is reached, and

retaining a last one f said low frequency information components.

17. A method of reconstructing an image from retained high frequency components of successive downsamplings and a low frequency component corresponding to a last one of said successively retained high frequency components, the method comprising:

combining said low frequency component with its corresponding high frequency component to form an initial combination,

iteratively enhancing said combination by enhancing step-like local behavior over said combination to form an enhanced combination,

using said enhanced combination together with a next one of said high frequency components to form a new initial combination,

repeating said steps of combining, and iteratively enhancing to form a new enhanced combination, and further forming new enhanced combination until all of said high frequency components are used, and

outputting a final enhanced combination as a reconstructed image.

18. Image enhancer for improving resolution of a pixilated image, comprising:

an upsampler for upsampling said image, and

an edge enhancer configured with a step function to scan said upsampled image, said step function being able to identify and enhance step like local behavior over said image, thereby to enhance resolution of said upsampled image.

19. Image storage apparatus for reducing the amount of storage capacity needed for storage of a given pixilated image, the apparatus comprising:

a downsampler for downsampling an image,

a low pass extractor associated with said downsampler for extracting a low pass component of said downsampled image, said low pass component being for storage,

an upsampler for upsampling said low pass component

and an enhancer configured with a step function to scan said upsampled image, said step function being able to identify and enhance step like local behavior over said image, thereby to restore said pixilated image.