Introduction
Recently, the CMOS image sensor (CIS) market crossed $11 billion per year. Its Compound Annual Growth Rate has been showing a double-digit growth each year because the digital camera has been widely used globally given that it is attached to almost all smartphones [1]. It is important to consider how best to utilize the small camera installed in the smartphone, which is not only used for taking pictures but also serves more important purposes. Si can perform photoelectric conversion from the ultraviolet to the near-infrared region. Digital still cameras usually have NIR cut filters, which allow the images to be viewed in visible light without acquiring an NIR wavelength [2]. However, a conventional digital camera with a common CIS can acquire a color image with only visible light (wavelength: 400 nm–700 nm). The near-infrared wavelength region is hardly utilized for photoelectric conversion. In these digital cameras, the near-infrared wavelength region is ignored [3]. Recently, some surveillance cameras could capture a black and white image with near-infrared (NIR) light, which uses only mono NIR wavelength [4]. Further, mono NIR wavelength is used in a 3D mapping sensor or in face recognition by KINECT [5]. Lately, in an industrial and a medical camera, multi-band spectral cameras are started to use. Such cameras use all band areas of the wavelength from ultraviolet to near-infrared [6]. However, the pixel size of the CIS should be designed to exceed approximately 5–10 μm square pixel size because of an inherent issue in the realization of spectral imaging technology [7].