Image processing method and apparatus

Abstract

For the purpose of conducting anisotropic diffusion filtering in a shorter time, an image processing method for conducting anisotropic diffusion filtering on a pixel value I i, j in a two-dimensional image, comprises: finding conduction coefficients C n , C s , C w , C e , C nw , C sw , C ne and C se in eight surrounding directions for each pixel based on a pixel value gradient ˆ‡I to produce a two-dimensional distribution image of the conduction coefficients; and finding pixel value first partial differentials ˆ‡ n I, ˆ‡ s I, ˆ‡ w I, ˆ‡ e I, ˆ‡ nw I, ˆ‡ sw I, ˆ‡ ne I and ˆ‡ se I in the eight surrounding directions for each pixel to calculate a pixel value in an output image according to: I i,j n +1 = I i,j n +»[[ C n �¢ ·ˆ‡ n I + C s �¢ ·ˆ‡ s I + C w �¢ ·ˆ‡ w I + C e �¢ ·ˆ‡ e I ] + 1 2 [ C nw �¢ ·ˆ‡ nw I + C sw �¢ ·ˆ‡ sw I + C ne �¢ ·ˆ‡ ne I + C nw �¢ ·ˆ‡ nw I ]] i,j n , (n: the number of times of repetition, »: a constant).

Description

technical field [0001] The present invention relates to an image processing method and device, more specifically to a method and device for implementing anisotropic diffusion filtering on a two-dimensional image. Background technique [0002] Filtering techniques capable of removing noise from a two-dimensional image without side effects such as edge degradation include anisotropic diffusion filtering. [0003] The anisotropic diffusion filtering includes obtaining pixel differential values ​​with respect to the nearest surrounding pixels of each pixel, obtaining a conductance coefficient (conduction coefficient) of each pixel based on the pixel value gradient, and using these conductance coefficients and pixel differential values ​​according to a predetermined Calculate the formula to calculate the pixel value in the output image. [0004] The calculation formula is of a space-proportional type in which the calculation result in one pass is used as an input for the next pa...

AI Technical Summary

Problems solved by technology

[0007] This anisotropic diffusion filtering consists of obtaining the pixel value gradient for each pixel and storing it once as a two-dimensional distribution image, reading the pixel value gradient from the two-dimensional distribution image to calculate the conductance coefficient, and then using this coefficient and the pixel differential value computes the pixel value in the output image, and performs this processing pixel by pixel, resulting in long and slow operations

Images