A novel approach for denoising and enhancement of extremely low-light video
Minjae Kim, Dubok Park, David K. Han, Hanseok Ko · IEEE Transactions on Consumer Electronics · 2015
In this paper, a novel approach for noise reduction and enhancement of extremely low-light video is proposed. For noise removal, a motion adaptive temporal filtering based on a Kalman structured updating is presented. Dynamic range of denoised video is increased by adjustment of RGB histograms using Gamma correction with adaptive clipping thresholds. Finally, residual noise is removed using a nonlocal means (NLM) denoising filter. The proposed method works directly on the color filter array (CFA) raw video for achieving low memory consumption1.