Advanced Impulse Detection & Reduction Based on Multimodal Filter

Abstract-Impulse noises are occurred in the images during image signal acquisition and processing from one location to another location. In this paper, the optimal detector noise filtering algorithm and its efficient hardware architecture is presented. The proposed architecture comprises of orthogonal direction pattern generation, sorter, thresholder, local binary converter, multimodal filter and pixel converter units respectively. The local binary converter unit detects and corrects the noise pixel efficiently using a simple logic circuit. The design possesses only two line memory buffers with very low computational complexity, thereby reducing the hardware cost and appropriate for several real-time applications.