Low-Light Image Enhancement Based on Quasi-Symmetric Correction Functions by Fusion

Abstract

Sometimes it is very difficult to obtain high-quality images because of the limitations of image-capturing devices and the environment. Gamma correction (GC) is widely used for image enhancement. However, traditional GC perhaps cannot preserve image details and may even reduce local contrast within high-illuminance regions. Therefore, we first define two couples of quasi-symmetric correction functions (QCFs) to solve these problems. Moreover, we propose a novel low-light image enhancement method based on proposed QCFs by fusion, which combines a globally-enhanced image by QCFs and a locally-enhanced image by contrast-limited adaptive histogram equalization (CLAHE). A large number of experimental results showed that our method could significantly enhance the detail and improve the contrast of low-light images. Our method also has a better performance than other state-of-the-art methods in both subjective and objective assessments.

SPLIE: Optimal Illumination Estimation for Structure Preserving Low-light Image Enhancement

“…Many image enhancement approaches have been introduced to enhance the brightness of low-light images, including histogram-based [1][2][3][4][5][6][7][8], deep learning-based [9][10][11][12][13][14][15][16], and Retinex-based methods [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37]. The simplest and more intuitive way to improve the visibility of weakly illuminated images is to increase their brightness.…”

“…Furthermore, Tan and Isa [7] introduced an adjusted HE enhancement algorithm which segments the original histogram into sub-histograms using histogram segmentation thresholds based on exposure regions. In [8], an enhancement algorithm was proposed depending on quasi-symmetric correction (QCFs) and histogram equalization. It combines the locally-enhanced image and the globally-enhanced image performed via contrast limited adaptive histogram equalization (CLAHE) [1] and QCFs, respectively.…”

“…, (8) where the symbols 𝛻 𝑣 and 𝛻 ℎ represent the first order derivatives over the vertical and horizontal directions, respectively. 𝐺 𝜎 (𝑥, 𝑦) denotes the Gaussian kernel with standard deviation 𝜎 and it is given by:…”

Detection of Road Images Containing a Counterlight Using Multilevel Analysis

“…Previous studies have applied intensity histogram equalization to increase the image intensity when it has been lowered by a counterlight. In general, the intensity histogram equalization method [20] assumes that the intensity value of the image except for the region including the counterlight has decreased because of the counterlight; uniformly applies the distribution of the intensity values. However, although the method is applied to improve the intensity values of pixels in a low-illuminance area, bright regions are excessively blurred because the pixel intensity values located in such regions also increase.…”