Device and method for compressing digital images

• US 20050213655A1
• Filed: 08/06/2004
• Published: 09/29/2005
• Est. Priority Date: 08/07/2003
• Status: Active Grant

• Current Assignee: InterDigital Madison Patent Holdings (InterDigital, Inc.)
• Original Assignee: Thomson Licensing (Vantiva SA)
• Inventors: Francois, Edouard, Guillotel, Philippe, Thoreau, Dominique, Bordes, Philippe
• Granted Patent: US 9,031,332 B2
• Time in Patent Office: Days
• US Class Current: 375/240.100
• CPC Class Codes: H04N 19/115 Selection of the code volume for a coding unit prior to coding
• CPC Class Codes: H04N 19/154 Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion use of rate-distortion criteria H04N19/147
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/34 Scalability techniques involving progressive bit-plane based encoding of the enhancement layer, e.g. fine granular scalability [FGS]

Abstract

The present invention relates to a device for compressing a source image by means of a base layer, coding a base image, and of an enhancement layer, coding a residual existing between the source image and the base image, the source image and the base image being structured as blocks of pixels. According to the invention, such a device is characterized in that it comprises means for generating the enhancement layer of a block as a function of the quality of restitution of this block, determined by comparing the coding of this block in the source image with the coding of this block in the base image, and/or as a function of information coded by this block of pixels.

Claims

1. Device for compressing a source image by means of a base layer, coding a base image, and of an enhancement layer, coding a residual existing between the source image and the base image, the source image and the base image being structured as blocks of pixels, comprising means for generating the enhancement layer of a block (k) as a function of the quality (psnr_k) of restitution of this block (k), determined by comparing the coding of this block in the source image with the coding of this block in the base image, and/or as a function of information coded by this block of pixels.

2. Device according to claim 1, comprising means for determining the quality (psnr_k) of restitution of the block (k) by means of a formula such as:

${\mathrm{psnr}}_k=-10<br><br>\mathrm{Log}<br><br>\left(\underset{i,j=0}{\overset{i,j=7}{\sum <br><br>}}<br><br>{\left({\mathrm{block}}_{k,\mathrm{source}}<br><br>\left(i,j\right)-{\mathrm{block}}_{,\mathrm{rec}}<br><br>\left(i,j\right)\right)}^2\right)/\left(L^2\times <br><br>64\right)$

where (i,j) represents the coordinates of a pixel of the block (k) considered, for example of dimensions 8×

8, in the source image (block_source) or in the base image (block,_rec) coded by the base layer, L representing the value of the maximum luminance of the image, for example equal to 255.

3. Device according to claim 1, comprising means for defining an average quality (Psnr_avg) of restitution of the image and for allocating a greater number of bits in the enhancement layer to the blocks whose restitution quality is less than this average quality (Pnsr_avg) or than a reference quality (Pnsr_ref) obtained on the basis of this average quality, for example by adding a quality margin.

4. Device according to claim 1, comprising means for comparing the quality (psnr_k) of restitution of the block k with a reference quality (psnr_ref) in such a way as to detect a deviation (δ

psnr_k) of quality with respect to this reference (psnr_ref) and means for allotting a number (NbBits_k) of bits of the enhancement layer to this block as a function of this deviation.

5. Device according to claim 4, comprising means for allotting a number (NbBits_k) of bits of the enhancement layer by using a proportionality relation such as:

${\mathrm{NbBits}}_k=\frac{\mathrm{\delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>{\mathrm{psnr}}_k}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{psnr}}\times <br><br>{\mathrm{throughput}}_{\mathrm{fgs}}$

where throughput_fgs is the quantity of bits per image contained in the train of the enhancement layer and Δ

psnr is a sum of deviations (δ

psnr_k) measured for the block considered.

6. Device according claim 1, comprising means for estimating the quality of restitution of the base layer by estimating the degradation, or the cost (CostFlti), of its filtering intended to reduce contrasts between neighbouring blocks and means for allotting to a block (k) of pixels a bigger number (NbBits_k) of bits of the enhancement layer the higher this cost (CostFlti).

7. Device according to claim 6, comprising means for estimating the cost (CostFlti) of the filtering with the aid of a formula such as:

${\mathrm{costFlt}}_i=\frac{\left(\underset{i,j=0}{\overset{i,j=7}{\sum <br><br>}}<br><br>\mathrm{abs}<br><br>\left({\mathrm{block}}_{\mathrm{baseFiltered}}^i<br><br>\left(i,j\right)-{\mathrm{block}}_{\mathrm{base}}^i<br><br>\left(i,j\right)\right)\right)}{64}$

where abs is the absolute value function, and block_basefiltered(i,j)−

block_base(i,j) represents the difference in values of the pixel with coordinates i and j of the base layer after filtering and before filtering.

8. Device according to claim 6, comprising means for generating the enhancement layer of a block as a function of a parameter relating to an average luminance (Lavg) associated with this block, determined for example by adding up the luminances of the pixels of the block considered.

9. Device according to claim 8, comprising means for weighting the cost of the filtering of a block as a function of a coefficient F(Lavg_i) whose value depends on the average luminance of this block.

10. Device according to claim 6, comprising means for weighting the cost of the filtering as a function of a coefficient F(Lavg_i) whose value depends on the texture, or activity, of the considered block of pixels.

11. Device according to claim 1, comprising means for generating the enhancement layer by considering a motion or displacement vector associated with each pixel block i, for example by means of the formula:

Vect _i ={square root}{square root over (Vx ² +Vy ² )}

where the components Vx and Vy represent the horizontal and vertical displacements of the block.

12. Device according to claim 1, comprising means for generating the enhancement layer by considering the possible exit of the block in an image subsequent to the image considered.

13. Device according to claim 1, comprising means for estimating the quality (psnr) of restitution of a block of pixels with the aid of an affine modelling such that the quality (psnr) of restitution of an image has a linear form:

psnr=Dpsnr·

x+b

where Dpsnr is a constant equal, for example, to 0.212 dB/bit per block and x represents the allotted number of bits per block.

14. Device according to claim 1, comprising means for generating the enhancement layer by allotting more bits to the blocks of an image if this image is used to generate other images, for example in the case of an inter-image predictive coding.

15. Device according to claim 1, comprising means for coding the base layer with the aid of a discrete cosine transform (DCT).

16. Method for compressing a source image by means of a base layer, coding a base image, and of an enhancement layer, coding a residual existing between the source image and the base image, the source image and the base image being structured as blocks of pixels, wherein the enhancement layer of a block is generated as a function of the quality (psnr_k) of restitution of this block (k), determined by comparing the coding of this block in the source image with the coding of this block in the base image, and/or as a function of information coded by this block of pixels.