Image processing apparatus, image processing method, program of image processing method, and recording medium in which program of image processing method has been recorded

• US 7,567,724 B2
• Filed: 04/21/2005
• Issued: 07/28/2009
• Est. Priority Date: 05/19/2004
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Ueda, Kazuhiko, Gomi, Shinichiro, Ogata, Masami
• Primary Examiner(s): Bali; Vikkram
• Assistant Examiner(s): ENTEZARI, MICHELLE M
• Application Number:
  US10/563,919
• Publication Number:
  US 20080123998A1
• Time in Patent Office: 1,559 Days
• Field of Search: 382/261, 382/269
• US Class Current: 382/261
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 3/403 Edge-driven scaling; Edge-based scaling
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 5/73 Deblurring; Sharpening
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 10/443 by matching or filtering
• CPC Class Codes: H04N 1/4092 Edge or detail enhancement

Abstract

The present invention is applied, for example, to resolution conversion. An edge gradient direction v1 with the largest gradient of pixel values and an edge direction v2 orthogonal to the edge gradient direction v1 are detected. Edge enhancement and smoothing processing are performed in the edge gradient direction v1 and in the edge direction v2, respectively, to generate output image data D2.

Claims

1. An image processing apparatus for processing input image data and for outputting output image data, the image processing apparatus comprising:

an edge detection unit for detecting, for pixels of the input image data, an edge gradient direction with the largest gradient of pixel values and an edge direction orthogonal to the edge gradient direction, setting a number of sampling points on a line in the edge direction, and generating sampling pixel values for the sampling points by interpolation;

an eigenvalue and eigenvector detection unit for detecting eigenvalues for the edge direction and the edge gradient direction, and calculating a reliability ratio of an edge in the edge direction with the eigenvalues;

an edge direction processing unit for smoothing the input image data in the edge direction, including the sampling pixel values, by filtering the pixel values, and outputting smoothed pixel values; and

an edge gradient direction processing unit for enhancing the input image data in the edge gradient direction and for outputting enhanced pixel values;

wherein the edge direction processing unit calculates a filtering range by multiplying the reliability ratio by an eigenvalue of the edge gradient direction, and changes a number of taps for the filtering based on the calculated filtering range.

2. The image processing apparatus according to claim 1, wherein the number of taps for the filtering is changed in a decimal fractional part by changing a weighting coefficient based on the reliability ratio, and by performing weighting addition of filtering results of different numbers of taps using the weighting coefficient.

3. The image processing apparatus according to claim 1, wherein the reliability ratio is a ratio of a dispersion of the gradient of the pixel values in the edge direction to a dispersion of the gradient of the pixel values in the edge gradient direction.

4. The image processing apparatus according to claim 1, wherein the edge gradient direction processing unit sets a number of gradient sampling points on a line in the edge gradient direction, generates gradient sampling pixel values for the gradient sampling points by interpolation, and performs filtering on the gradient sampling pixel values.

5. The image processing apparatus according to claim 1,

wherein the output image data is image data obtained by changing a sampling pitch of the input image data, and

wherein the image processing apparatus further includes:

an interpolation processing unit for performing interpolation on the input image data and for outputting interpolated image data with a sampling pitch of the output image data;

a blend ratio determination unit for changing a weighting coefficient for blending based on the reliability ratio; and

a blend processing unit for performing weighting addition of the image data output from the edge gradient direction processing unit and the interpolated image data using the weighting coefficient for blending, and for outputting the output image data.

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

a blend ratio determination unit for changing a weighting coefficient for blending based on the reliability ratio of an edge in the edge direction; and

a blend processing unit for performing weighting addition of the enhanced pixel values and the input image data using the weighting coefficient for blending, and for outputting the output image data.

7. An image processing method for processing input image data and for outputting output image data, the image processing method comprising:

using a computer to perform the steps of

detecting, for pixels of the input image data, an edge gradient direction with the largest gradient of pixel values and an edge direction orthogonal to the edge gradient direction for;

setting a number of sampling points on a line in the edge direction;

generating sampling pixel values for the sampling points by interpolation;

detecting eigenvalues for the edge direction and the edge gradient direction, and calculating a reliability ratio of an edge in the edge direction with the eigenvalues;

smoothing the input image data in the edge direction, including the sampling pixel values, by filtering the pixel values and outputting smoothed pixel values;

enhancing the input image data in the edge gradient direction and outputting enhanced pixel values;

calculating a filtering range by multiplying the reliability ratio by an eigenvalue of the edge gradient direction; and

changing a number of taps for the filtering based on the calculated filtering range.

8. A computer-readable recording medium storing a program for an image processing method performed by arithmetic processing means for processing input image data and for outputting output image data, the program for the image processing method comprising:

detecting, for pixels of the input image data, an edge gradient direction with the largest gradient of pixel values and an edge direction orthogonal to the edge gradient direction for;

setting a number of sampling points on a line in the edge direction;

generating sampling pixel values for the sampling points by interpolation;

detecting eigenvalues for the edge direction and the edge gradient direction, and calculating a reliability ratio of an edge in the edge direction with the eigenvalues;

smoothing the input image data in the edge direction, including the sampling pixel values, by filtering the pixel values and outputting smoothed pixel values;

enhancing the input image data in the edge gradient direction and outputting enhanced pixel values;

calculating a filtering range by multiplying the reliability ratio by an eigenvalue of the edge gradient direction; and

changing a number of taps for the filtering based on the calculated filtering range.