A Simple Yet Robust Nonlinear Function for Low-Light Image Enhancement Task

Abstract

We present a novel, parameter-free nonlinear transformation for low-light image enhancement that operates directly on individual pixel values. This simple yet powerful function requires no prior knowledge or external tuning, and enhances image brightness and contrast by leveraging only the input image itself. When applied iteratively, the method achieves optimal results after just three applications. Despite its minimalism, our approach outperforms recent state-of-the-art methods on benchmarks. This highlights the potential of simple signal processing operations for emergent enhancement, and suggests directions for theoretical analysis, integration with deep learning, and deployment in real-world vision systems.