Adaptive trimmed averaging filter for noise removal in color images

Abstract

In the paper a novel filtering design intended for the removal of mixed impulsive and Gaussian noise contaminating color images is proposed. The new denoising scheme is based on the concept of the robustified vector median, which determines the center of a cluster of most similar pixels. The filter output is the average of pixels belonging to the cluster and thus it can be treated as an adaptive trimmed averaging filter. The new filtering design is capable of suppressing mixed impulsive and Gaussian noise, while preserving and even enhancing image edges. The experiments revealed that the new noise reduction design outperforms the standard trimmed vector median and other state-of-the-art methods. The simplicity of the proposed filter and its low computational load enables its applications in real time image and video enhancement applications.