초록 ▼

First similarity values along at least four directions are ascertained within a local area containing a target pixel and weighted averaging is performed by adding the pixel values of pixels around the target pixel value to the pixel value of the target pixel, adding weight along a direction having a

First similarity values along at least four directions are ascertained within a local area containing a target pixel and weighted averaging is performed by adding the pixel values of pixels around the target pixel value to the pixel value of the target pixel, adding weight along a direction having a small first similarity value (along a direction manifesting a high degree of similarity). By incorporating the pixel value level differences among a plurality of pixels on adjacent lines extending adjacent to the target pixel into the first similarity values, it becomes possible to effectively remove jaggies that are difficult to eliminate in the prior art. Furthermore, by making a judgment on degrees of similarity by incorporating color information such as characteristics differences among different color pixels, a more accurate judgment can be made with regard to the image structure to enable very accurate direction-dependent low-pass filtering.

대표청구항 ▼

What is claimed is: 1. An image processing method for implementing low-pass filtering on image data, comprising: a similarity judging step in which similarity among pixels are judged along four directions in a local area containing a target pixel undergoing low-pass filtering processing; and a dire

What is claimed is: 1. An image processing method for implementing low-pass filtering on image data, comprising: a similarity judging step in which similarity among pixels are judged along four directions in a local area containing a target pixel undergoing low-pass filtering processing; and a direction-dependent low-pass filtering step of performing a weighted averaging operation in which weighted pixel values of pixels around a target pixel are added to a pixel value of said target pixel and a result of said addition is divided by a sum of the weights, a weighting rate along a direction manifesting marked similarity becoming increased based upon said judgment obtained in said similarity judging step, wherein: in said similarity judging step, first similarity values tt1, yy1, nu1 and ns1 along four directions are calculated using absolute values of differences among pixel values of a plurality of pixels along each direction within said local area; a representative value m1 is calculated by averaging or taking a median of said first similarity values tt1, yy1, nu1 and ns1 along four directions; second similarity values tt2, yy2, nu2 and ns2 are calculated based upon said first similarity values tt1, yy1, nu1 and ns1 along four directions and said representative value m1, where at least one of second similarity values tt2, yy2, nu2 and ns2 is 0 except in a case in which said first similarity values tt1, yy1, nu1 and ns1 are same, and as said second similarity values tt2, yy2, nu2 and ns2 become larger, a degree of similarity among pixels increases, and in said direction-dependent low-pass filtering step, weighting rates for neighboring pixels located along four directions are determined in correspondence to said calculated second similarity values tt2, yy2, nu2 and ns2, the pixel values of pixels around said target pixel are weighted by said weighting rates. 2. An image processing method according to claim 1, wherein: said second similarity values tt2, yy2, nu2 and ns2 are calculated using expressions 1-4: description="In-line Formulae" end="lead"tt2=max{m1-tt1+δ, γ} (expression 1)description="In-line Formulae" end="tail" description="In-line Formulae" end="lead"yy2=max{m1-y1+δ, γ} (expression 2)description="In-line Formulae" end="tail" description="In-line Formulae" end="lead"nu2=max{m1-nu1+δ, γ} (expression 3)description="In-line Formulae" end="tail" description="In-line Formulae" end="lead"ns2=max{m1-ns1+δ, γ} (expression 4)description="In-line Formulae" end="tail" (where δ and γ in the expressions above each represents a predetermined value which may be 0). 3. An image processing method according to claim 1, wherein: said image data are color image data; and said similarity is judged based upon at least two types of color information in said color image data in said similarity judging step. 4. An image processing method according to claim 3, wherein: said similarity is judged based upon color image data yet to undergo interpolation processing in said similarity judging step. 5. An image processing method according to claim 1, wherein: said image data are image data having undergone interpolation processing to interpolate pixels with missing color components; and said low-pass filtering processing is implemented only on target pixels having undergone said interpolation processing in said direction-dependent low-pass filtering step. 6. An image processing method according to claim 1, wherein: said image data are image data having undergone interpolation processing to interpolate pixels with missing color components; and a pixel value of each pixel having undergone said interpolation processing is limited by a threshold value corresponding to a largest pixel value or a smallest pixel value in a specific area near the corresponding pixel prior to the low-pass filtering processing in the similarity judging step. 7. An image processing method according to claim 1, wherein: said image data are color image data having, at least, a first color with a highest pixel density and a second color with a low pixel density and vacancies of color information, said image processing method further comprising: a color difference calculating step in which a color difference between said second color and said first color is obtained for each pixel at which said second color is present; a color difference interpolating step in which a color difference interpolation value is obtained for a pixel at which said second color is not present based upon said color difference obtained in said color difference calculating step; and a second color restoring step in which said second color is restored based upon said color difference interpolation value obtained in said color difference interpolating step and a pixel value of said first color, wherein: said first color used to calculate said color difference in said color difference calculating step is said first color that has not undergone said low-pass filtering processing. 8. An image processing method according to claim 7, wherein: said first color used in restoring said second color in said second color restoring step is said first color that has undergone said low-pass filtering processing. 9. An image processing method according to claim 1, wherein: in said direction-dependent low-pass filtering step, a first weighting rate, a second weighting rate, a third weighting rate, a fourth weighting rate and a fifth weighting rate are respectively applied for said target pixel, pixels above and below said target pixel, pixels next to said target pixel on the right and left, pixel above said target pixel on the right and pixel below said target pixel on the left, pixel above said target pixel on the left and pixel below said target pixel on the right; and said second, third, fourth and fifth weighting rates along a direction manifesting marked similarity becomes increased. 10. An image processing method according to claim 9, wherein: said first weighting rate is 1/(1+2k), said second weighting rate is k횞tt/(1+2k), said third weighting rate is k횞yy/(1+2k), said fourth weighting rate is k X nu/(1+2k) and said fifth weighting rate is k횞ns/(1+ 2k), where k represents a predetermined value, and tt, yy, nu and ns satisfy the following equation: tt+yy+nu+ns=1.