Adaptive directional edge-preserving false contour filtering

Abstract

A method for adaptive false contour reduction includes: detecting contour location information by removing a flat region having certain brightness values from a first input image based on a bright value difference between the first input image and a second input image with a bit depth of the first input image reduced; detecting false-contour direction and location information by measuring directional contrast of the flat region-removed first image, and distinguishing the false contour area and an edge area out of the false contour location information based on the measured contrast; and smoothing the false contour area by using the false contour direction and location information, and removing the false contour from the first input image. Thus, a flat region can be automatically removed by using a brightness value difference between a bit depth-reduced image and an original input image and detect a false contour, thereby enhancing a precision degree of false contour detection. Further, signal components can be prevented from being degraded by performing smoothing over only a false contour.