Low-Light Video Enhancement Using Improved Illumination Map and Retinex Theory

Low-light video enhancement is crucial in applications such as surveillance, automotive systems, and medical imaging, where visibility is affected by poor light conditions. This project proposes a new method for low-light video enhancement based on the Retinex theory. The Retinex algorithm, which is widely used for image enhancement, divides an image into illuminant and reflectance components, simulating human visual perception. However, when applied to low-light videos, the standard Retinex algorithm faces challenges such as noise amplification and temporal anomalies.