IMAGE PROCESSING APPARATUS, IMAGE PROCESSING METHOD, AND PROGRAM

• US 20110150356A1
• Filed: 12/14/2010
• Published: 06/23/2011
• Est. Priority Date: 12/22/2009
• Status: Active Grant

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: JO, Kensei, Nomura, Yoshikuni, Mitsunaga, Tomoo
• Granted Patent: US 8,411,991 B2
• Time in Patent Office: Days
• US Class Current: 382/269
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 5/90 Dynamic range modification of images or parts thereof

Abstract

An image processing apparatus includes: a characteristic value calculation unit calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks; and an edge-preserving smoothing unit calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block containing noticed pixels of the input image, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels.

Claims

1. An image processing apparatus comprising:

characteristic value calculation means for calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks;

edge-preserving smoothing means for calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels;

tone curve calculation means for calculating a shape of a tone curve for each of the pixels of the input image based on the general luminance value of the pixels of the input image; and

luminance value gray-scale correction means for correcting a gray scale of the luminance values of each of the pixels of the input image using the tone curve for each of the pixels of the input image.

2. The image processing apparatus according to claim 1,

wherein the weight function is the Gauss function, and

wherein the edge-preserving smoothing means calculates the general luminance value of the noticed pixels using a linear function in which a value based on a dispersion of the weight function and the block luminance dispersion of the noticed block is set as a slope and a value based on the slope and the block luminance average value of the noticed block is set as an intercept.

3. The image processing apparatus according to claim 2, wherein on the assumption that the dispersion of the weight function is Va, the block luminance dispersion of the noticed block is Vb, the block luminance average value of the noticed block is u, the slope is a, the intercept is b, the luminance value of the noticed pixel is L, and the general luminance value of the noticed pixels is Ll, the linear function is Ll=a×

L+b, where a=Vb÷

(Va+Vb) and b=(1−

a)×

u are satisfied.

4. The image processing apparatus according to claim 2,

wherein the characteristic value calculation means further calculates the slope and the intercept of the linear function for each spatial block,

wherein the image processing apparatus further comprises:

maintenance means for maintaining the slope and the intercept of the linear function for each spatial block only during a period corresponding to one frame of the input image, and

wherein the edge-preserving smoothing means calculates the general luminance value of the pixels of the input image based on the slope and the intercept of the linear function for each spatial block maintained by the maintenance means.

5. The image processing apparatus according to claim 1, wherein the edge-preserving smoothing means calculates the general luminance value of the noticed pixels by calculating the general luminance values in neighboring blocks which are a predetermined number of spatial blocks in the vicinity of the noticed pixels, using the block luminance average value and the block luminance dispersion of each neighboring block and the weight function for the noticed pixels and by weight-averaging the general luminance value of each neighboring block based on a distance between the neighboring block and the noticed pixel.

6. The image processing apparatus according to claim 1, further comprising:

reduced-image generation means for generating a reduced image formed by reducing the input image,

wherein the characteristic value calculation means calculates the block luminance average value and the block luminance dispersion of each of spatial blocks obtained by dividing the reduced image in the space direction, and

wherein the edge-preserving smoothing means calculates the general luminance value of the pixels of the input image using the block luminance average values and the block luminance dispersions of the reduced image.

7. The image processing apparatus according to claim 1,

wherein the characteristic value calculation means spatially smoothes the block luminance average value and the block luminance dispersion of each spatial block, and

wherein the edge-preserving smoothing means calculates the general luminance value of the pixels of the input image using the block luminance average value and the block luminance dispersion spatially smoothed.

8. The image processing apparatus according to claim 1, wherein when the general luminance value is subjected to gray-scale correction by the tone curve, the tone curve calculation means calculates the shape of the tone curve so that the general luminance value is converted into substantially a middle value of a range in which the luminance values subjected to the gray-scale correction by the tone curve is acquired.

9. The image processing apparatus according to claim 1, further comprising:

general luminance value gray-scale correction means for correcting the gray scale of the general luminance value of the pixels of the input image using the tone curve for each of the pixels of the input image; and

contrast correction means for correcting a contrast of the luminance value subjected to the gray-scale correction by the luminance value gray-scale correction means based on the general luminance value subjected to the gray-scale correction by the general luminance value gray-scale correction means and the shape of the tone curve.

10. An image processing method performed by an image processing apparatus performing gray-scale correction of an image, the image processing method comprising the steps of:

calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks;

calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels;

calculating a shape of a tone curve for each of the pixels of the input image based on the general luminance value of the pixels of the input image; and

correcting a gray scale of the luminance values of each of the pixels of the input image using the tone curve for each of the pixels of the input image.

11. A program causing a computer to execute the steps of:

calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks;

calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels;

calculating a shape of a tone curve for each of the pixels of the input image based on the general luminance value of the pixels of the input image; and

correcting a gray scale of the luminance values of each of the pixels of the input image using the tone curve for each of the pixels of the input image.

12. An image processing apparatus comprising:

characteristic value calculation means for calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks; and

edge-preserving smoothing means for calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels.

13. The image processing apparatus according to claim 12,

wherein the weight function is the Gauss function, and

wherein the edge-preserving smoothing means calculates the general luminance value of the noticed pixels using a linear function in which a value based on a dispersion of the weight function and the block luminance dispersion of the noticed block is set as a slope and a value based on the slope and a block luminance average value of the noticed block is set as an intercept.

14. The image processing apparatus according to claim 13, wherein on the assumption that the dispersion of the weight function is Va, the block luminance dispersion of the noticed block is Vb, the block luminance average value of the noticed block is u, the slope is a, the intercept is b, the luminance value of the noticed pixel is L, and the general luminance value of the noticed pixels is Ll, the linear function is Ll=a×

L+b, where a=Vb÷

(Va+Vb) and b=(1−

a)×

u are satisfied.

15. The image processing apparatus according to claim 12, wherein the edge-preserving smoothing means calculates the general luminance value of the noticed pixels by calculating the general luminance values in neighboring blocks which are a predetermined number of spatial blocks in the vicinity of the noticed pixels, using the block luminance average value and the block luminance dispersion of each neighboring block and the weight function for the noticed pixels and by weight-averaging the general luminance value of each neighboring block based on a distance between the neighboring block and the noticed pixel.

16. An image processing method performed by an image processing apparatus performing an edge-preserving smoothing process on an image, the image processing method comprising the steps of:

calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks; and

calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels.

17. A program causing a computer to execute the steps of:

calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks; and

calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels.

18. An image processing apparatus comprising:

a characteristic value calculation unit calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks;

an edge-preserving smoothing unit calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels;

a tone curve calculation unit calculating a shape of a tone curve for each of the pixels of the input image based on the general luminance value of the pixels of the input image; and

a luminance value gray-scale correction unit correcting a gray scale of the luminance values of each of the pixels of the input image using the tone curve for each of the pixels of the input image.

19. An image processing apparatus comprising:

a characteristic value calculation unit calculating a block luminance average value, which is an average value of luminance values in each of spatial blocks obtained by dividing an input image in a space direction, and a block luminance dispersion, which is a dispersion of the luminance values in each of the spatial blocks; and

an edge-preserving smoothing unit calculating a general luminance value of pixels of the input image by setting a value obtained by approximately calculating a weighted average value of the luminance values of a noticed block which is the spatial block to which noticed pixels of the input image belong, on the assumption that a distribution of the luminance values of each of the spatial blocks conforms to a Gaussian distribution based on the block luminance average value and the block luminance dispersion of each of the spatial blocks, as the general luminance value of the noticed pixels, the weighted average value being calculated using a weight function, in which a weight of the luminance values increases as the luminance values are closer to the luminance values of the noticed pixels.