FC_ICM及其在图像增强中的应用  

FC_ICM and Its Application in Image Enhancement

作　　者：杨忠斌[1] 吕明钊[2] 胡宝雷[1] 

机构地区：[1]中国航空工业西安航空计算技术研究所第1研究室,陕西西安710068 [2]中国航空工业西安航空计算技术研究所质量部,陕西西安710068

出　　处：《电子科技》2013年第7期67-71,73,共6页Electronic Science and Technology

摘　　要：提出了一种改进的可连续点火的交叉视觉皮质模型,并将其运用到图像增强当中。新模型利用了点火阈值对神经元行为的长期记忆能力,将其作为反馈和外部激励共同决定神经元的状态值;将阈值衰减因子替换为函数变换,从而可根据具体应用灵活选择合适的衰减函数;同时在点火神经元的阈值提升时增加了邻域判读,以决定对应阈值是否发生陡变,且实现了神经元的连续点火。在此基础上,提出了基于FC_ICM的直方图均衡化方法,对比基于CLAHE和ICM的图像增强方法,实验结果表明,该方法可有效提升全局和局部对比度,及有效抑制噪声。An improved continuous ignition intersecting cortical model is presented, which is applied in image enhancement. The new model makes use of the ignition threshold neuron behavior of long-term memory capacity as their feedback combined with external excitation neurons to decide the status value ; the threshold attenuation factor is replaced by the transformation of the function, thus a flexibility to select the appropriate attenuation function depen- ding on the applications; the interpretation of the neighborhood is added in the ignition threshold neurons enhance- ment in order to determine whether the corresponding threshold sudden changes occur, and to realize continuous igni- tion of the neurons. On this basis, the histogram equalization method is put forward based on FC_ICM in contrast to CLAHE and ICM image enhancement methods. Experimental results show that our method can effectively enhance the global and local contrast and effectively suppress noise.

关 键 词：交叉视觉皮质模型 图像增强 可连续点火 直方图均衡 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]