Abstract
To improve the observability of low-light images, a low-light image enhancement algorithm based on nonsubsampled shearlet transform (NSST) is presented (LIEST). The proposed algorithm can synchronously achieve contrast improvement, noise suppression, and the enhancement of specific directional details. An enhancement framework of low-light noisy images is first derived, and then, according to the framework, a low-light noisy image is decomposed into low-pass subband coefficients and bandpass direction subband coefficients by NSST. Then, in the NSST domain, an illumination map is estimated based on a bright channel of the low-pass subband coefficients, and noise is simultaneously suppressed by shrinking the bandpass direction subband coefficients. Finally, based on the estimated illumination map, the low-pass subband coefficients, and the shrunken bandpass direction subband coefficients, inverse NSST is implemented to achieve low-light image enhancement. Experiments demonstrate that the LIEST exhibits superior performance in improving contrast, suppressing noise, and highlighting specific details as compared to seven similar algorithms.
Highlights
Low-light conditions often occur during photo shooting, such as while photographing at night time or taking pictures under trees or tunnels, etc
Low-light image enhancement [1] plays an important role in image preprocessing, as its performance directly affects the success of image processing in the steps, including image segmentation, object detection, and image classification [2], [3], etc
Hao et al [25], [26] presented two algorithms to enhance low-light images based on the simple Retinex model; the method presented in their earlier study improves the contrast of images in different illumination conditions, and the bright channel of the image refined by a Gaussian filter was used as the illumination of the simple Retinex model to enhance low-light images in the later study
Summary
Low-light conditions often occur during photo shooting, such as while photographing at night time or taking pictures under trees or tunnels, etc. Hao et al [25], [26] presented two algorithms to enhance low-light images based on the simple Retinex model; the method presented in their earlier study improves the contrast of images in different illumination conditions, and the bright channel of the image refined by a Gaussian filter was used as the illumination of the simple Retinex model to enhance low-light images in the later study. Guo et al [29] optimized the illumination map by constraining the L1 norm of the gradient of the illumination map, enhanced a low-light image according to the simple Retinex model, and suppressed the noise of the enhanced image using the BM3D denoising algorithm. According to the preceding analysis, this work presents a low-light image enhancement algorithm based on nonsubsampled shearlet transform (LIEST).
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have