Method of monitoring the quality of distributed digital images by detecting false contours

• US 6,707,943 B2
• Filed: 02/05/2001
• Issued: 03/16/2004
• Est. Priority Date: 02/21/2000
• Status: Active Grant

• Current Assignee: TDF (Fluor Corporation)
• Original Assignee: Orange S.A., Telediffusion De France SA
• Inventors: Wyckens, Emmanuel, Pefferkohn, Stéphane, Blin, Jean-Louis, Gicquel, Jean-Charles
• Primary Examiner(s): Johnson, Timothy M.
• Assistant Examiner(s): Kassa, Yosef
• Application Number:
  US09/777,028
• Publication Number:
  US 20010022852A1
• Time in Patent Office: 1,135 Days
• Field of Search: 382/165, 382/107, 382/190, 382/199, 382/218, 382/219, 382/242, 382/266, 382/274, 358/537, 358/538, 358/452, 358/453, 348/26, 348/29, 348/14.15, 348/96, 348/97, 348/99, 348/100, 348/108
• US Class Current: 382/199
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/30168 Image quality inspection
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection

Abstract

A method and device for monitoring the quality of distributed digital images coded by blocks of pixels in which a false contour phenomenon may be generated when the image is reproduced. An image average for pixels represented by at least one luminance respectively chrominance components is calculated and a false contour effect on the basis of a criterion for discriminating the luminance or chrominance component difference between adjacent pixels of adjacent groups of pixels is detected in at least one reference direction of each current image. A visibility coefficient of at least one current image is calculated from the value of the image average speed vector and psycho-visual criteria relating to the existence of the false contour effect in the reference direction.

Claims

1. A method of monitoring the quality of digital images coded by blocks of pixels in which a false contour phenomenon is generated while decoding and reconstructing said coded images, wherein said method consists of, for each successive current image:

calculating an image average speed vector representative of the average speed of pixels represented by at least one of their luminance, respectively chrominance, components, between said current image and an image preceding said current image;

detecting in at least one reference direction of said current image a false contour effect on the basis of a criterion for discriminating the luminance, respectively the chrominance component difference between adjacent pixels of adjacent groups of pixels; and

calculating a visibility coefficient of at least one current image from the value of said image average speed vector of said current image and psycho-visual criteria relating to the existence of said false contour effect in said reference direction.

2. The method of claim 1, wherein said at least one reference direction of said current image is the horizontal line scanning direction and/or the vertical frame scanning direction.

3. The method of claim 1, wherein said step of calculating an average speed vector consists of:

dividing said current image into adjacent image blocks each defined by a particular number of lines and columns of image pixels;

calculating for each block of said current image a luminance, respectively chrominance, average component value, said average component value being defined by a line component vector representative of the average of the components of successive image pixels of columns constituting said image block and a column component vector representative of the average of the components of pixels of successive images of lines constituting the same image block;

calculating a block movement vector for each block from a block movement vector component in said first reference direction and a block movement vector component in said second reference direction, each block movement vector component in said first respectively said second reference direction being established as the distance expressed as a number of pixels in each direction of a group of pixels whose luminance respectively chrominance component difference is a minimum; and

calculating said image average speed vector as the average of said block movement vectors for all the blocks constituting said current image.

4. The method of claim 1, wherein said step of detecting a false contour effect in at least one reference direction of said current image consists of at least:

calculating the absolute value of the difference of the luminance respectively the chrominance components between the adjacent pixels of each pair of successive rows of pixels in said reference direction to constitute a difference image in said reference direction; and

generating a binary image representative of the existence of potential false contours in said reference direction from said difference image and assigning each pixel a particular binary value if the absolute value of the difference is greater than the absolute values of the difference of the luminance respectively the chrominance components of a number h of adjacent rows of pixels in the same reference direction or otherwise a complemented binary value.

5. The method of claim 4, wherein said step of detecting a false contour effect in at least one reference direction of said current image further includes, after obtaining said binary image representative of the existence of potential false contours in said reference direction, a step of eliminating from said binary image spurious potential false contours to generate a final binary image representative of false contours, the binary value of the pixels of said final binary image being established so that:

each point of said final binary image is assigned the binary value of the corresponding point of said binary image representative of the existence of potential false contours if said binary value is the complemented binary value representative of the absence of a potential false contour; and

for any point in a row in the reference direction of the binary image representative of a potential false contour and whose binary value is the binary value representative of a potential false contour, the final binary value assigned representative of the existence or the absence of a final false contour is determined by an interpretation procedure.

6. The method of claim 5, wherein said interpretation procedure consists in:

discriminating a number l of successive points whose binary value is equal to said binary value representative of a potential false contour over said row in the reference direction and in a parallel row offset from said row by a number k of pixels; and

comparing said number l of consecutive points with a threshold value L representative of the existence of a false contour, and

if the number l of consecutive points is less than said threshold value L, assigning to any point of said final binary image the complemented binary value representative of the absence of potential false contours, said potential false contour corresponding to a spurious potential false contour being thus eliminated;

if the number l of consecutive points is greater than said threshold value L;

calculating the signal-to-noise ratio of each segment of points contained in an area of said binary image representative of potential false contours, said area being defined as an individual block whose apex corresponds to said point and whose dimension in the first respectively the second reference direction is L×

h;

comparing said signal-to-noise ratio to a particular value and, if said signal-to-noise ratio is greater than said particular value;

assigning the binary value representative of a false contour to any point of said final binary image belonging to the segment of length L in said first reference direction and to any point of said final binary image belonging to a parallel segment of the same length offset by the amount k in said second reference direction if said signal-to-noise ratio is greater than said threshold value, and

assigning the complemented binary value representative of the absence of false contours to said points belonging to said segments of said final binary image otherwise.

7. The method of claim 4, wherein said step of detecting a false contour effect in at least one reference direction of said current image further consists in calculating for said binary image corresponding to said current image a coefficient representative of the existence of false contours by summing all the binary values representative of a false contour.

8. The method according to claim 1, wherein, for a particular sequence of a set of successive current images, said method further consists in:

calculating the overall average speed for said particular set of successive current images as the average of the average speeds of each successive current image; and

calculating for said particular set of successive current images an average false contour effect defined as the average of the coefficients representative of the existence of false contours of each successive current image.

9. The method of claim 1, wherein said psycho-visual criteria consist in determining the visibility of the average false contour effect vis-à

-vis a particular threshold value, said average false contour effect being declared visible when its value is greater than said particular threshold value and invisible otherwise.

10. The method of claim 9, wherein said particular threshold value is a linear combination of said overall average speed with the parameters of said linear combination being determined experimentally.

11. A device for monitoring the quality of digital images coded by blocks of pixels in which a false contour phenomenon is generated while decoding and reconstructing said coded image, said device including in combination:

a module for converting digital images into a dedicated format including separate luminance and chrominance components;

a module for acquiring and separating luminance respectively chrominance components of successive images;

a module receiving said luminance respectively chrominance components for detecting false contour effects in at least one reference direction on the basis of a criterion for discriminating said luminance respectively said chrominance component difference between adjacent pixels of adjacent groups of pixels from the value of an image average speed vector representative of the average speed of pixels represented by said luminance respectively said chrominance components; and

a module for calculating and displaying at least one visibility coefficient of at least one current image from the value of said image average speed vector of said current image and psycho-visual criteria relating to the existence of said false contour effect in said reference direction.

12. The device of claim 11, further including a module for calculating complementary parameters such as the bit rate of the digital signal conveying said digital images.