摘要:
面向非均匀光照图像, 提出了基于向心自动波交叉皮质模型(Centripetal-autowave intersecting cortical model, CA-ICM)的图像增强算法. 为了解决原始交叉皮质模型(Intersecting cortical model, ICM)固有自动波效应在图像增强应用中易导致边缘模糊的问题, 首先,设计了基于形态学中值集的向心自动波(Centripetal autowave, CA)实现方式. 提出了基于图像特征---键值(Key)的自适应S形状映射函数, 以此作为CA-ICM模型的输入输出的映射关系. 为了增强算法的鲁棒性, 对未点火位置进行了标注和修复. 最后提出了非线性变换的颜色恢复方法. 同时对模型参数设计进行了细致讨论. 仿真结果表明, 该模型可以有效进行光照动态范围的调整, 向心自动波约束产生了邻域内的侧抑制作用, 输出图像对比度得到大幅提升, 细节边缘清晰, 颜色恢复充分自然, 客观评价值高.
Abstract:
For non-uniform-lighting images, we propose an image enhancement technique based on centripetal-autowave intersecting cortical model (CA-ICM). The original ICM possesses the autowave nature stemming from the connection function during the firing process, but poses a problem called interference, which could blur the edge and detail in image processing tasks. To solve it, the implementation of CA based on morphological median set is presented. As for the relationship between the input and output of CA-ICM, we apply an adaptive S shape non-linear mapping function based on image characteristics of key value. Furthermore, we label and restore those unfired positions for algorithm's robustness. A modified non-linear color restoration process based on chromatic information is applied finally. The precise mapping function and optimized parameters considered in detail lead to better experimental results, indicating that the CA-ICM shows more advantages than ICM. Efficient dynamic range adjustment is conducted, especially the highlight inhibition and shadow rendition with details. The CA creates the lateral inhibition effect, leading to nature, sharp, and colorful outputs with high objective evaluations.