A New Structural Non-liner Image Decomposition Type Image processing and its Quality Control for High-Quality and High-Performance Imaging.

用数字彩色相机观测的颜色数据经历由光学低通滤光器(抗混叠滤光器)引起的空间退化、由于其光电转换中的信号相关的散粒噪声而引起的点退化、利用拜耳的原色滤色器阵列进行颜色马赛克等。从观测到的噪声模糊的彩色数据中恢复丢失的信息将使生成高质量的彩色图像成为可能，因此图像恢复是数字彩色相机图像处理中的关键问题。研究如何从不完全观测的图像数据中恢复出高质量的彩色图像。我们设计了各种图像处理技术，并通过计算机模拟对其性能进行了评估。

Color data observed with a digital color camera undergo spatial degradations caused by an optical low-pass filter(anti-alias filter), point degradations due to signal-dependent shot noise in its photoelectric conversion, color-mosaicking with the Bayer's primary color filter array, and so on. The restoration of lost information from the observed noisy blurry color data will make it possible to produce a high-quality color image, and thus the image restoration is a key problem in image-processing of a digital color camera. This research deals with how to restore a high-quality color image from the incompletely observed image data. We devised various image processing techniques, and evaluated its performance by computer simulations.