High-ISO image de-noising using burst filtering
Nour Aburaed, Harish Bhaskar, Faisal Shah Khan · 2016
Cameras allow capturing images in low light conditions by adjusting the level of their ISO setting. ISO, which stands for International Standards Organization, controls the sensitivity of a camera's sensor. A high-ISO value amplifies not only the image's signal but also the noise signal, giving both of them a high gain depending on the chosen ISO level. This results in statistically unpredictable noise distribution with different intensities all over the image. In this paper, this type of high-ISO noise and its different characteristics will be investigated; particularly its effect in the luminance and chrominance channels. Based on these differences, a burst filtering algorithm that is based on bilateral filtering and median stacking is proposed.