With the continuous advancement of digitalization and intelligent technologies, acquiring clear digital images under low-light conditions has emerged as a crucial research direction in the field of image processing. Imaging under dim lighting environments typically suffers from significant degradation in luminance and contrast, coupled with severe noise contamination, making denoising processing for such images increasingly vital. This paper systematically reviews current mainstream low-light image denoising algorithms, categorizing them into two primary classes: filter-based methods and deep learning-based approaches. Through comprehensive analyses of theoretical frameworks and experimental evaluations, representative algorithms from both categories are critically examined and compared. Furthermore, this study prospectively discusses future research directions, highlighting potential technical pathways including performance stabilization in complex scenarios, construction of lightweight models, and exploration of unsupervised or self-supervised methodologies. The primary objective of this work is to provide researchers in related domains with methodological references and technical insights through systematic analysis and forward-looking perspectives.
Dong et al. (Mon,) studied this question.