Image quality enhancing apparatus, image display apparatus, image quality enhancing method, and computer readable storage medium

Patent metadata

Field	Value
Publication number	US-9836821-B2
Application number	US-201615139330-A
Country	US
Kind code	B2
Filing date	Apr 27, 2016
Priority date	Nov 1, 2013
Publication date	Dec 5, 2017
Grant date	Dec 5, 2017

Abstract

An image quality enhancing apparatus which make a learning-type image quality enhancing method utilizing a sparse expression practical are provided. The apparatus calculates, from the feature quantity of an image, coefficients of low-image-quality base vectors expressing a feature quantity with a linear sum and generates the image with the image quality enhanced by calculating a linear sum of high-image-quality base vectors using the calculated coefficient. When calculating the coefficient, the number of base vectors with non-zero coefficients is determined, the determined number of base vectors is selected, and a solution of a coefficient matrix is calculated by assuming the coefficients of the base vectors other than the selected base vectors are zero. The amount of processes necessary for obtaining a sparse solution of a coefficient matrix can be reduced by adjusting the number of base vectors with non-zero coefficients, and a practical image quality enhancing apparatus can be realized.

First claim

What is claimed is:

1. An image quality enhancing apparatus comprising:
a section which stores a plurality of first feature quantities which can express a feature quantity in any image with a linear sum thereof;
a section which stores a plurality of second feature quantities which indicate a feature quantity of a high image quality corresponding to the first feature quantities;
a coefficient calculating section which calculates, from a feature quantity of an image targeted for image quality enhancement, a coefficient for expressing the feature quantity with the linear sum of the plurality of first feature quantities; and
a section which generates an enhanced image quality image which is the image with the image quality enhanced by replacing each of the first feature quantities included in the linear sum using the coefficient calculated by the coefficient calculating section with corresponding second feature quantities; wherein
the coefficient calculating section comprises:
a determining section which determines the number of first feature quantities to be used for calculation among the plurality of first feature quantities based on a correspondence relationship between a feature quantity of any image and the number of first feature quantities necessary for expressing the feature quantity with a linear sum of fewer first feature quantities than the total number of the plurality of first feature quantities;
a selecting section which selects the number of first feature quantities which has been determined by the determining section from among the plurality of first feature quantities; and
a section which calculates the coefficient based on the feature quantity and the first feature quantities selected by the selecting section.

2. The image quality enhancing apparatus according to claim 1 , wherein the section which calculates the coefficient lets, among the coefficient for expressing the feature quantity with the linear sum, a coefficient to be multiplied by the first feature quantities other than the first feature quantities selected by the selecting section be zero, and calculates a coefficient to be multiplied by the first feature quantities selected by the selecting section.

3. The image quality enhancing apparatus according to claim 1 , further comprising a correspondence relationship storage section which stores the correspondence relationship between the feature quantity of any image and the number of the first feature quantities necessary for expressing the feature quantity with a linear sum of fewer first feature quantities than the total number of the plurality of first feature quantities, wherein
the determining section is configured to determine the number of first feature quantities corresponding to the feature quantity of the image targeted for image quality enhancement based on the correspondence relationship.

4. The image quality enhancing apparatus according to claim 3 , wherein the correspondence relationship storage section stores the correspondence relationship by amplitude of noise included in any image,
and the determining section comprises:
a section which calculates the amplitude of the noise included in the image targeted for image quality enhancement;
a section which identifies the correspondence relationship according to the amplitude of the noise calculated by the section which calculates the amplitude of the noise; and
a section which determines the number of first feature quantities corresponding to the feature quantity of the image based on the correspondence relationship identified by the section which identifies the correspondence relationship.

5. The image quality enhancing apparatus according to claim 4 , wherein the correspondence relationship storage section stores a correspondence relationship such that the larger the amplitude of noise, the fewer the number of first feature quantities corresponding to feature quantities of any image becomes.

6. The image quality enhancing apparatus according to claim 1 further comprising:
a section which generates the feature quantity of the image by extracting a high frequency component that is higher than a predetermined spatial frequency from among a spatial frequency component of a luminance distribution in the image targeted for image quality enhancement;
a section which extracts a low frequency component of a lower frequency than the high frequency component among the spatial frequency component when the number determined by the determining section is larger than zero;
a section which generates an image with an enhanced image quality by adding a low frequency component extracted by the section which extracts a low frequency component and the feature quantity of an image with an enhanced image quality;
a section which removes noise included in the image targeted for image quality enhancement by an edge-preservation filter when the number determined by the determining section is zero; and
a section which handles an image after the noise is removed by the section which removes noise as an image with an enhanced image quality.

7. The image quality enhancing apparatus according to claim 1 further comprising:
a section which calculates an amplification factor equal to or larger than zero and equal to or smaller than 1 which, based on a histogram indicating the relationship between a value of a predetermined noise included in each pixel of any image and a pixel count having the value, approaches zero as a pixel count in the histogram in accordance with the value of the predetermined noise corresponding to a value in each pixel of the feature quantity of the image targeted for image quality enhancement becomes larger, and approaches 1 as the pixel count becomes smaller; and
a section which multiplies the amplification factor calculated by the section which calculates the amplification factor with the value in each pixel of the feature quantity of the image targeted for image quality enhancement.

8. An image display apparatus comprising:
the image quality enhancing apparatus according to claim 1 ; and
a section which displays an image whose image quality enhanced by the image quality enhancing apparatus.

9. An image quality enhancing method comprising:
storing a plurality of first feature quantities that can express a feature quantity in any image with a linear sum thereof;
storing a plurality of second feature quantities indicating a feature quantity of a high image quality corresponding to the first feature quantities;
calculating, from the feature quantity of an image targeted for image quality enhancement, a coefficient for expressing the feature quantity with the linear sum of the plurality of first feature quantities; and
generating an image which is the image with the image quality enhanced by replacing each of the first feature quantities included in the linear sum which has used the calculated coefficient with corresponding second feature quantities, wherein the image quality enhancing method comprises:
determining, based on a correspondence relationship between a feature quantity of any image and the number of first feature quantities necessary for expressing the feature quantity with a linear sum of fewer first feature quantities than the total number of the plurality of first feature quantities, the number of first feature quantities used for calculation among the plurality of first feature quantities;
selecting the number of first feature quantities which has been determined, from among the plurality of first feature quantities; and
calculating the coefficient based on the selected first feature quantities and the feature quantity.

10. The image quality enhancing method according to claim 9 , wherein the

Assignees

• Eizo Corp

Inventors

• Higashi Masafumi
• Chen Haifeng
• Aoki Reo
• Matsuzaki Hiroki
• Katagiri Tadayoshi

Classifications

• G06T3/4076 Primary

  using the original low-resolution images to iteratively correct the high-resolution images · CPC title

• G06T5/60

  using machine learning, e.g. neural networks · CPC title

• G06T3/4053 Primary

  based on super-resolution, i.e. the output image resolution being higher than the sensor resolution · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• H04N7/0117

  involving conversion of the spatial resolution of the incoming video signal (for graphics images G09G2340/0407) · CPC title