Intensity mapping function based weighted frame averaging for high dynamic range imaging

TLDR: In this article, a weighted frame averaging method based on an intensity mapping function (IMF) is proposed to reduce noise from the input low dynamic range (LDR) images with shorter exposures.

Abstract: A high quality image can be synthesized by combining several differently exposed low dynamic range (LDR) images of the same scene. For scenes under low lighting condition, cameras are usually set to high sensitivity mode to reduce exposure times and avoid motion blur on captured images. However, those images tend to be noisy and the noise severely degrades the visual quality of final image especially on dark areas. In this paper, a weighted frame averaging method based on an intensity mapping function (IMF) is proposed to reduce noise from the input LDR images with shorter exposures. The proposed method does not require any knowledge on either camera response functions (CRFs) or exposure times, and it is much simpler than the state-of-art method. Experiment results also show that the proposed method effectively removes the noise from the shortly exposed LDR images without introducing any blurring or other artifacts.