有效滤除高强度图像混合噪声的方法

摘要

传统的交叉视觉皮质模型(ICM)对单一噪声的去除具有良好的性能.为了扩展ICM在图像降噪领域的应用,提高降噪能力,提出一种基于邻域连接的NL-ICM.针对传统ICM存在的局限性,在神经元的构造上引入双边滤波的思想,通过扩展神经元的连接输入、引入连接权重、设计脉冲阈值实时计算函数,并为神经元设计像素更新规则.实验结果表明,该模型能够较好地去除图像中的混合噪声.%Intersecting cortical model (ICM) is only capable of filtering images with only one single type of noise. In order to extend the application of ICM in image denoising, ICM neurons' connection is re-designed. In our work, the thought of Bilateral Filtering is introduced together with extending the connecting input of neurons. By considering the linking-weight, designing a real-time pulse threshold function and a pixel renewal rule, a new Neighborhood-Linking ICM is proposed in this paper. Experiments show that the proposed ICM-based filtering method is fast and effective for removing the impulse noises mixed with additional Gaussian noises.