关键词: 人工智能, 图像去噪, 交叉视觉皮层模型, 赋时矩阵

Abstract: To overcome the problem of multi-sensor image de-noising, an adaptive technique for image de-noising based on Improved Intersecting Cortical Model (IICM) is proposed. To begin with, the basic classic ICM is analyzed and its structural drawbacks are detected, and IICM is presented, which not only declines the number of parameters required setting, but constructs the time matrix T to settle the problem of setting the number of iterative times, with which the transformation from the image spatial information to time information can be realized. Subsequently, the noisy pixels are located by utilizing the information provided by the time matrix T in IICM. Finally, the task of efficiently filtering the impulsive noise in the image is completed by the proposed adaptive algorithm for image de-noising based on IICM. Owing to dealing only with the polluted pixels and adjusting the size of area window adaptively, the proposed technique has remarkable superiority over other ones in both simulated performance and running efficiency.

Key words: artificial intelligence, image de-noising, intersecting cortical model, time matrix