K. Hwang and T.H. Nishimura (Japan)
Noise reduction, CMOS image sensor, 3D filter
We propose a weighted filter with 3D structure for 1/f noise reduction in complementary metal oxide semiconductor image sensors (CMOS Image Sensors: CIS) by applying characteristics of noise distribution on time domain. The algorithms using time domain or image stacking that focus on same pixel position are proposed to make up for the weakness points of general method on edge region. In the case of using self-pixel position, some problems are occurred such as blocking effect or difficulty of acquisition of images. The proposed algorithm use multi-images which are generated artificially through the noise distribution. It is optimized by using weight value on the center pixel of each 2D mask. The purpose of proposed method is to obtain the best PSNR with minimum number of images and minimum weight value. We used White Gaussian noise distribution that 1/f noise has. Through the proposed algorithm, we have obtained the better PSNR than other methods as a mean filter, Wiener filter and self pixel based method. The algorithm is expected that it is effective for other random noise with other types of noise as photon shot noise with Poisson distribution.
Important Links:
Go Back
