Local tone mapping for HDR video

1. A tone mapping method of an image processing device, the method comprising:

dividing a first image of a succession of images into a plurality of sub-blocks of pixels;

applying a logarithmic function to an intensity of each pixel of the first image;

generating a first mini-image comprising pixels, each pixel of the first mini-image representing a corresponding sub-block of the first image, an intensity of each pixel of the first mini-image being representative of an intensity of the pixels of the corresponding sub-block after application of the logarithmic function to the pixels of the corresponding sub-block;

storing, the first mini-image in a memory; and

modifying a second image of the succession of images based on the first mini-image, generating an output image, wherein the modifying the second image comprises applying, to each pixel of the second image, a gain value GMP q according to:

GMP q =(LOCAL_GAIN q −1)*GLOBAL_POWER+1,

where GLOBAL_POWER is a value representing an influence of the modification on the whole image and LOCAL_GAIN q is a local contrast gain, computed based on a logarithmic representation of the second image and the mini-image of the first image.

2. The method of claim 1 , wherein the modifying comprises, for each pixel of the second image:

computing of a context value, the context value depending on:

a spatial proximity between the pixel of the second image and centers of sub-blocks of the second image adjacent to a sub-block of the second image comprising the pixel of the second image; and

an affinity in intensity between the pixel of the second image and pixels of the first mini-image; and

using the context value to generate the local contrast gain.

3. The method of claim 2 , wherein each pixel OUTPUT_P q of the output image is equal to OUTPUT_P q =GMP q *P q , where P q represents a value of a channel of the pixel, and OUTPUT_P q represents a value of the corresponding channel in the output image.

4. The method of claim 1 , comprising storing the output image in the memory or another memory or displaying the output image.

5. The method of claim 1 , wherein each sub-block comprises at least two pixels.

6. The method of claim 1 , wherein the intensity of each pixel of the first mini-image is computed based on an average of the intensities of the pixels of the corresponding sub-block.

7. The method of claim 1 , wherein the intensity of each pixel of the first mini-image is computed based on:

an average of the intensities of the pixels in a square centered on the corresponding sub-block, a size of the square being equal to k times the size of the sub-block, k being a number between 0 and 2, not including 0; or

an average of the intensities of the pixels in a circle centered on the corresponding sub-block, a radius of the circle being equal to k times the size of the sub-block, k being a number between 0 and 2, not including 0.

8. The method of claim 1 , wherein the first mini-image used for the generation of the output image is erased from the memory after the generation of the output image.

9. The method of claim 1 , wherein a respective mini-image is generated for all the images of the succession of images.

10. The method of claim 2 , wherein the modification of the pixel of the second image is independent of the context value if:

the intensity of the pixel is above a first threshold, or below a second threshold;

the pixel is a color pixel and the intensity of one of the colors of the pixel is above a third threshold;

an accuracy of the context value is below a fourth threshold; or

combinations thereof.

11. A device, comprising:

a memory; and

image processing circuitry coupled to the memory, wherein the image processing circuitry, in operation:

divides a first image of a succession of images into a plurality of sub-blocks of pixels;

applies a logarithmic function to an intensity of each pixel of the first image;

generates a first mini-image comprising pixels, each pixel of the first mini-image representing a corresponding sub-block of the first image, an intensity of each pixel of the first mini-image being representative of an intensity of the pixels of the corresponding sub-block after application of the logarithmic function to the pixels of the corresponding sub-block;

stores the first mini-image; and

applies tone mapping to a second image of the succession of images based on the first mini-image, generating an output image, wherein the applying tone mapping comprises applying, to each pixel of the second image, a gain value GMP q according to:

GMP q =(LOCAL_GAIN q −1)*GLOBAL_POWER+1,

where GLOBAL_POWER is a value representing an influence of the tone mapping on the whole image and LOCAL_GAIN q is a local contrast gain, computed based on a logarithmic representation of the second image and the mini-image of the first image.

12. The device of claim 11 , wherein the image processing circuitry, in operation, for each pixel of the second image:

computes a context value, the context value depending on:

a spatial proximity between the pixel of the second image and centers of sub-blocks of the second image adjacent to a sub-block of the second image comprising the pixel of the second image; and

an affinity in intensity between the pixel of the second image and pixels of the first mini-image; and

uses the context value to generate a value of the local contrast gain.

13. The device of claim 12 , wherein each pixel OUTPUT_P q of the output image is equal to OUTPUT_P q =GMP q *P q , where P q represents a value of a channel of the pixel, and OUTPUT_P q represents a value of the corresponding channel in the output image.

14. The device of claim 11 , wherein each sub-block comprises a plurality of pixels.

15. The device of claim 11 , wherein the intensity of each pixel of the first mini-image is computed based on an average of the intensities of the pixels of the corresponding sub-block.

16. The device of claim 11 , wherein the intensity of each pixel of the first mini-image is computed based on:

an average of the intensities of the pixels in a square centered on the corresponding sub-block, a size of the square being equal to k times the size of the sub-block, k being a number between 0 and 2, not including 0; or

an average of the intensities of the pixels in a circle centered on the corresponding sub-block, a radius of the circle being equal to k times the size of the sub-block, k being a number between 0 and 2, not including 0.

17. The device of claim 11 , wherein the image processing circuitry, in operation, replaces the first mini-image used for the generation of the output image with a mini-image of the second image after the generation of the output image.

18. The device of claim 11 , wherein the image processing circuitry, in operation, successively generates a respective mini-image for the images of the succession of images.

19. A system, comprising:

a display, which, in operation, displays images; and

image processing circuitry coupled to the display, wherein the image processing circuitry, in operation:

divides a first image of a succession of images into a plurality of sub-blocks of pixels;

applies a logarithmic function to an intensity of each pixel of the first image;

generates a first mini-image comprising pixels, each pixel of the first mini-image representing a corresponding sub-block of the first image, an intensity of each pixel of the first mini-image being representative of an intensity of the pixels of the corresponding sub-block after application of the logarithmic function to the pixels of the corresponding sub-block;

stores the first mini-image; and

applies tone mapping to a second image of the succession of images based on the first mini-image, generating an output image, wherein the applying tone mapping comprises applying, to each pixel of the second image, a gain value GMP q according to:

GMP q =(LOCAL_GAIN q −1)*GLOBAL_POWER+1,

where GLOBAL_POWER is a value representing an influence of the tone mapping on the whole image and LOCAL_GAIN q is a local contrast gain, computed based on a logarithmic representation of the second image and the mini-image of the first image.

20. The system of claim 19 , wherein the display, in operation, displays the output image.

21. The system of claim 19 , wherein the image processing circuitry, in operation, for each pixel of the second image:

computes a context value, the context value depending on:

a spatial proximity between the pixel of the second image and centers of sub-blocks of the second image adjacent to a sub-block of the second image comprising the pixel of the second image; and

an affinity in intensity between the pixel of the second image and the pixels of the first mini-image; and

uses the context value to generate a value of the local contrast gain.

22. A non-transitory computer-readable medium having contents which cause image processing circuitry to perform a tone mapping method, the tone mapping method comprising:

dividing a first image of a succession of images into a plurality of sub-blocks of pixels;

applying a logarithmic function to an intensity of each pixel of the first image;

generating a first mini-image comprising pixels, each pixel of the first mini-image representing a corresponding sub-block of the first image, an intensity of each pixel of the first mini-image being representative of an intensity of the pixels of the corresponding sub-block after application of the logarithmic function to the pixels of the corresponding sub-block;

storing, the first mini-image in a memory; and

modifying a second image of the succession of images based on the first mini-image, generating an output image, wherein the modifying comprises applying, to each pixel of the second image, a gain value GMP q according to:

GMP q =(LOCAL_GAIN q −1)*GLOBAL_POWER+1,

where GLOBAL_POWER is a value representing an influence of the modifying on the whole image and LOCAL_GAIN q is a local contrast gain, computed based on a logarithmic representation of the second image and the mini-image of the first image.

23. The non-transitory computer-readable medium of claim 22 , wherein the modifying comprises, for each pixel of the second image:

computing of a context value, the context value depending on:

a spatial proximity between the pixel of the second image and centers of sub-blocks of the second image adjacent to a sub-block of the second image comprising the pixel of the second image; and

an affinity in intensity between the pixel of the second image and pixels of the first mini-image; and

using the context value to generate a value of the local contrast gain.

24. The non-transitory computer-readable medium of claim 22 , wherein the contents comprise instructions executed by the image processing circuitry.