LIME: Low Light Image Enhancement Via Illumination Map Estimation

Low-light images often suffer from poor visibility, reducing their visual quality and adversely affecting various computer vision and multimedia applications. This paper presents a robust low-light image enhancement method, known as Low-light Image Enhancement, which effectively enhances image brightness and contrast while preserving structural details. The approach first estimates the illumination of each pixel by extracting the maximum intensity among the R, G, and B channels. A refined illumination map is then generated using a structure-prior-based regularization technique. To further improve the enhancement quality, a guided filter is applied to refine the transmission map, ensuring smooth illumination adjustments. The enhanced luminance values are mapped using a gamma correction function, followed by local adaptation using kernel filtering. Additionally, BM3D denoising is employed to suppress noise, improving overall image clarity. The proposed method is evaluated using various performance metrics, including Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Lightness Order Error (LOE). Experimental results demonstrate that LIME outperforms existing state-of-the-art methods in both visual quality and computational efficiency, making it a promising solution for real-world low-light image enhancement applications.

Keywords: Image Processing Steps, Enhancement, luminance map and PSNR.