System and method for enhancement of coded images using adaptive spatial filtering

• US 5,812,702 A
• Filed: 11/22/1995
• Issued: 09/22/1998
• Est. Priority Date: 11/22/1995
• Status: Expired due to Term

• Current Assignee: Comcast MO Group Incorporated (Comcast Corporation)
• Original Assignee: U S West Inc.
• Inventors: Kundu, Amlan
• Primary Examiner(s): Boureau, Leo H.
• Assistant Examiner(s): Mariam, Daniel G.
• Application Number:
  US08/561,679
• Time in Patent Office: 1,035 Days
• Field of Search: 382/261, 382/262, 382/231, 382/233, 382/235, 382/260
• US Class Current: 382/260
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 7/231 using full search

Abstract

Pixels in a low bit-rate encoded image are categorized into two or more categories. In one embodiment three pixel categories are used: 1) pixels belonging to quasi-constant (QC) regions where the pixel intensity values vary slowly (e.g., pictorial portions), 2) pixels belonging to textured regions which are characterized by many small edges and thin-line signals (e.g., text portions), and 3) pixels belonging to dominant-edge (DE) regions which are characterized by few sharp and dominant edges (e.g., edge portions). In one embodiment DE pixels are categorized first, then QC and textural pixels are distinguished from the remaining pixels using the number of zero-crossings among the pixels and a threshold. Conventional spatial filters that are well suited for each pixel category type are then used to enhance each region of the image. In one embodiment, various combinations of spatial filters are used to enhance the image.

Claims

1. A method for enhancing a low bit-rate encoded image, the method implemented on a computer having a memory and comprising the steps of:

receiving a low bit-rate encoded image comprising pixels;

classifying said pixels into a first, a second, and a third pixel category, wherein the first, second, and third pixel categories are quasi-constant, textural, and dominant edge, respectively, and wherein the step of classifying comprises the substeps of;

setting the value of selected pixels in the received low bit-rate encoded image to zero to form a modified image;

determining an a-level zero-crossing measurement as a function of the pixels in the modified image and the pixels in the received low bit-rate encoded image; and

using the a-level zero-crossing measurement to classify said pixels into the quasi-constant and textural categories;

applying a first enhancement method to the pixel in the quasi-constant category; and

applying a second enhancement method to the pixels in the dominant-edge category;

whereby an enhanced image is generated.

2. The method of claim 1, wherein the low bit-rate encoded image is selected from the group consisting of a JPEG encoded image, a low-bit vector quantization encoded image, and a low-bit fractal encoded image.

3. The method of claim 1, wherein each of said pixels is classified into one of the first, second, and third pixel catagories.

4. The method of claim 1, wherein a pixel is classified into two or more of the pixel categories.

5. The method of claim 1, wherein the quasi-constant category comprises pixels, the textural category comprises texels, and the dominant edge category comprises edgels.

6. The method of claim 1, wherein the step of classifying comprises the steps of:

determining pixels belonging to the dominant edge category from said pixels in the received low bit-rate encoded image; and

determining pixels in the quasi-constant category and the textural category from remaining pixels in the received low bit-rate encoded image.

7. The method of claim 1, wherein the step of classifying furether comprises the substeps of:

determining a zero-crossing measurement as a function of the pixels in the received low bit-rate encoded image; and

comparing the a-level zero-crossing measurement and the zero-crossing measurement;

wherein, the substep of using the a-level zero-crossing measurement to classify said pixels is performed as a function of the comparing substep.

8. The method of claim 1, wherein the step of classifying further comprises the substeps of:

determining a threshold value;

wherein, the substep of using the a-level zero-crossing measurement to classify said pixels is performed as a function of the a-level zero-crossing measurement and the threshold.

9. The method of claim 1, wherein the first enhancement method comprises the step of:

enhancing pixels in the quasi-constant category using a thresholded D-filter.

10. The method of claim 1, wherein the second enhancement method comprises the step of:

enhancing pixels in the dominant edge category using a median filter.

11. The method of claim 1, wherein the second enhancement method comprises the step of:

enhancing pixels in the dominant edge category using a multi-stage median filter.

12. The method of claim 1, wherein:

the second enhancement method comprises enhancing pixels in the dominant edge category using a multi-stage median filter to produce enhanced dominat edge pixels;

the first enhancement method comprises enhancing the quasi-constant category using a thresholded D-filter to produce enhanced quasi-constant pixels and

the method further comprises the step of enhancing the enhanced dominat edge pixels, the enhanced quasi-constant pixels, and the pixels in the textural category using a thresholded D-filter.

15. The method of claim 1, the method further comprising the step of applying a third enhancement method to the pixels in the textural category.

13. Computer memory programmed to enhance a low bit-rate encoded image, the program comprising the steps of:

receiving a low bit-rate encoded image comprising pixels;

classifying said pixels into a first, a second, and a third pixel category, wherein the first, second and third pixel categories are quasi-constant, textural, and dominant edge, respectively, wherein the step of classifying comprises the substeps of;

setting the value of selected pixels in the received low bit-rate encoded image to zero to form a modified image;

determining an a-level zero-crossing measurement as a function of the pixels in the modified image of the pixels in the received low bit-rate encoded image;

determining a threshold value; and

classifying pixels in the quasi-constant category and pixels in the textural category as a function of the a-level zero-crossing measurement and the threshold;

applying a first enhancement method to the pixel in the quasi-constant category; and

applying a second enhancement method to the pixels in the dominant-edge category;

whereby an enhanced image is generated.

14. A computer storage device having a program for enhancing a low bit-rate encoded image, the program comprising the steps of:

receiving a low bit-rate encoded image comprising pixels;

classifying said pixels into a first, a second, and a third pixel category, wherein the first, second, and third pixel category, wherein the first, second and third pixel categories are quasi-constant, textural, and dominant edge, respectively, wherein the step of classifying comprises the substeps of;

determining a zero-crossing measurement as a function of the pixels in the reveived low bit-rate encoded image;

setting the value of selected pixels in the received low bit-rate encoded image to zero to form a modified image;

determining an a-level zero-crossing measurement as a function of the pixels in the modified image of the pixels in the received low bit-rate encoded image;

comparing the zero-crossing measurement and the a-level zero-crossing measurement; and

classifying pixels in the quasi-constant category and pixels in the textural category as a function of the comparing substep;

applying a first enhancement method to the pixels in the quasi-constant category; and

applying a second enhancement method to the pixels in the dominant edge category;

whereby an enhanced image is generated.