Abstract
The visual quality of photographs taken under imperfect lightness conditions can be degenerated by multiple factors, e.g., low lightness, imaging noise, color distortion, and so on. Current low-light image enhancement models focus on the improvement of low lightness only, or simply deal with all the degeneration factors as a whole, therefore leading to sub-optimal results. In this article, we propose to decouple the enhancement model into two sequential stages. The first stage focuses on improving the scene visibility based on a pixel-wise non-linear mapping. The second stage focuses on improving the appearance fidelity by suppressing the rest degeneration factors. The decoupled model facilitates the enhancement in two aspects. On the one hand, the whole low-light enhancement can be divided into two easier subtasks. The first one only aims to enhance the visibility. It also helps to bridge the large intensity gap between the low-light and normal-light images. In this way, the second subtask can be described as the local appearance adjustment. On the other hand, since the parameter matrix learned from the first stage is aware of the lightness distribution and the scene structure, it can be incorporated into the second stage as the complementary information. In the experiments, our model demonstrates the state-of-the-art performance in both qualitative and quantitative comparisons, compared with other low-light image enhancement models. In addition, the ablation studies also validate the effectiveness of our model in multiple aspects, such as model structure and loss function.
- [1] . 2016. Layer normalization. arXiv preprint arXiv:1607.06450 (2016).Google Scholar
- [2] . 2017. A joint intrinsic-extrinsic prior model for retinex. In Proceedings of the International Conference on Computer Vision (ICCV’17). IEEE, 4020–4029.
DOI: Google ScholarCross Ref
- [3] . 2018. Learning a deep single image contrast enhancer from multi-exposure images. IEEE Trans. Image Process. 27, 4 (
Apr. 2018), 2049–2062.DOI: Google ScholarCross Ref
- [4] . 2011. Fast efficient algorithm for enhancement of low lighting video. In Proceedings of the IEEE International Conference on Multimedia and Expo (ICME’11). IEEE, 1–6.
DOI: Google ScholarDigital Library
- [5] . 2016. A fusion-based enhancing method for weakly illuminated images. Signal Process. 129 (
Dec. 2016), 82–96.DOI: Google ScholarDigital Library
- [6] . 2016. A weighted variational model for simultaneous reflectance and illumination estimation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’16). IEEE, 2782–2790.
DOI: Google ScholarCross Ref
- [7] . 2015. Using free energy principle for blind image quality assessment. IEEE Trans. Multimedia 17, 1 (
Jan. 2015), 50–63.DOI: Google ScholarCross Ref
- [8] . 2020. Zero-reference deep curve estimation for low-light image enhancement. In Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR’2020). IEEE, 1777–1786.
DOI: Google ScholarCross Ref
- [9] . 2017. LIME: Low-light image enhancement via illumination map estimation. IEEE Trans. Image Process. 26, 2 (
Feb. 2017), 982–993.DOI: Google ScholarDigital Library
- [10] . 2019. Lightness-aware contrast enhancement for images with different illumination conditions. Multimedia Tools Appl. 78, 3 (
Feb. 2019), 3817–3830. Google ScholarDigital Library
- [11] . 2020. Low-light image enhancement with semi-decoupled decomposition. IEEE Trans. Multimedia 22, 12 (
Dec. 2020), 3025–3038.DOI: Google ScholarDigital Library
- [12] . 2015. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. In Proceedings of the IEEE International Conference on Computer Vision (ICCV’15). IEEE, 1026–1034.
DOI: Google ScholarDigital Library
- [13] . 2017. FC4: Fully convolutional color constancy with confidence-weighted pooling. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’17). IEEE, 330–339.
DOI: Google ScholarCross Ref
- [14] . 2015. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In Proceedings of the International Conference on Machine Learning (ICML’15). JMLR.org, 448–456.Google Scholar
- [15] . 2021. EnlightenGAN: Deep light enhancement without paired supervision. IEEE Trans. Image Process. 30 (
Jan. 2021), 2340–2349.DOI: Google ScholarDigital Library
- [16] . 2017. Photo-realistic single image super-resolution using a generative adversarial network. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’17). IEEE, 105–114.
DOI: Google ScholarCross Ref
- [17] . 2012. Contrast enhancement based on layered difference representation. In Proceedings of the IEEE International Conference on Image Processing (ICIP’12). IEEE, 965–968.
DOI: Google ScholarCross Ref
- [18] . 2013. Contrast enhancement based on layered difference representation of 2D histograms. IEEE Trans. Image Process. 22, 12 (
Dec. 2013), 5372–5384.DOI: Google ScholarDigital Library
- [19] . 2021. Low-light image and video enhancement using deep learning: A survey. (2021).
arxiv:cs.CV/2104.10729 Google Scholar - [20] . 2021. Generative adversarial network for low-light image enhancement. IET Image Process. (2021).Google Scholar
- [21] . 2018. Structure-revealing low-light image enhancement via robust retinex model. IEEE Trans. Image Process. 27, 6 (
Jun. 2018), 2828–2841.Google Scholar - [22] . 2021. AMBCR: Low-light image enhancement via attention guided multi-branch construction and Retinex theory. IET Image Process. (2021).Google Scholar
- [23] . 2014. Single image layer separation using relative smoothness. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’14). IEEE, 2752–2759.
DOI: Google ScholarDigital Library
- [24] . 2017. LLNet: A deep autoencoder approach to natural low-light image enhancement. Pattern Recognit. 61 (
Jan. 2017), 650–662.DOI: Google ScholarDigital Library
- [25] . 2017. Robust multi-exposure image fusion: A structural patch decomposition approach. IEEE Trans. Image Process. 26, 5 (
May. 2017), 2519–2532.DOI: Google ScholarDigital Library
- [26] . 2011. Blind/referenceless image spatial quality evaluator. In Proceedings of the Asilomar Conference on Signals, Systems and Computers (ACSCC’11). IEEE, 723–727.
DOI: Google ScholarCross Ref
- [27] . 2013. Making a “completely blind” image quality analyzer. IEEE Signal Process. Lett. 20, 3 (
May. 2013), 209–212.DOI: Google ScholarCross Ref
- [28] . 2018. Joint enhancement and denoising method via sequential decomposition. In Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS’18). 1–5.
DOI: Google ScholarCross Ref
- [29] . 2020. LR3M: Robust low-light enhancement via low-rank regularized Retinex model. IEEE Trans. Image Process. 29, 12 (
Apr. 2020), 5862–5876.DOI: Google ScholarCross Ref
- [30] . 2004. Realization of the contrast limited adaptive histogram equalization (CLAHE) for real-time image enhancement. J. VLSI Signal Process. 38, 1 (
Aug. 2004), 35–44.Google ScholarCross Ref
- [31] . 2015. U-net: Convolutional networks for biomedical image segmentation. In Proceedings of the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI’15). Springer, 234–241.
DOI: Google ScholarCross Ref
- [32] . 2011. Recovering intrinsic images with a global sparsity prior on reflectance. In Proceedings of the Conference on Neural Information Processing Systems (NeurIPS’11). 765–773.Google Scholar
- [33] . 2021. A low-light image enhancement method based on bright channel prior and maximum colour channel. IET Image Process. (2021).Google Scholar
Cross Ref
- [34] . 2015. Very deep convolutional networks for large-scale image recognition. In Proceedings of the International Conference on Learning Representations (ICLR’15).Google Scholar
- [35] . 2015. Blind image quality evaluation using perception based features. In Proceedings of the IEEE National Conference on Communications (NCC’15). IEEE, 1–6.
DOI: Google ScholarCross Ref
- [36] . 2008. Fast centre-surround contrast modification. IET Image Process. 2, 1 (
Feb. 2008), 19–34.DOI: Google ScholarCross Ref
- [37] . 2016. A fusion-based method for single backlit image enhancement. In Proceedings of the IEEE International Conference on Image Processing (ICIP’16). IEEE, 4077–4081.
DOI: Google ScholarCross Ref
- [38] . 2019. Underexposed photo enhancement using deep illumination estimation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’19). IEEE, 6842–6850.
DOI: Google ScholarCross Ref
- [39] . 2018. GLADNet: Low-light enhancement network with global awareness. In Proceedings of the IEEE International Conference on Automatic Face & Gesture Recognition (FG’18). IEEE, 751–755.
DOI: Google ScholarDigital Library
- [40] . 2004. Image quality assessment: From error visibility to structural similarity. IEEE Trans. Image Process. 13, 4 (
Apr. 2004), 600–612.DOI: Google ScholarDigital Library
- [41] . 2018. Deep Retinex decomposition for low-light enhancement. In Proceedings of the British Machine Vision Conference (BMVC’18). BMVA Press, 1–12.Google Scholar
- [42] . 2020. From fidelity to perceptual quality: A semi-supervised approach for low-light image enhancement. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR’2020). IEEE, 3063–3072.
DOI: Google ScholarCross Ref
- [43] . 2021. Band representation-based semi-supervised low-light image enhancement: Bridging the gap between signal fidelity and perceptual quality. IEEE Trans. Image Process. 30 (2021), 3461–3473.
DOI: Google ScholarDigital Library
- [44] . 2018. High-quality exposure correction of underexposed photos. In Proceedings of the 26th ACM International Conference on Multimedia (ACM MM’18). ACM, 582–590.
DOI: Google ScholarDigital Library
- [45] . 2019. Kindling the darkness: A practical low-light image enhancer. In Proceedings of the 27th ACM International Conference on Multimedia (ACM MM’19). ACM, 1632–1640.
DOI: Google ScholarDigital Library
Index Terms
Decoupled Low-Light Image Enhancement
Recommendations
Adaptive Variational Model for Contrast Enhancement of Low-Light Images
Contrast enhancement plays an important role in image/video processing and computer vision applications. Its main purpose is to adjust the image intensity to enhance the quality and features of the image. In this paper, we propose a simple and efficient ...
Low-light image enhancement based on membership function and gamma correction
AbstractThe aim of low-light image enhancement algorithms is to improve the luminance of images. However, existing low-light image enhancement algorithms inevitably cause an enhanced image to be over- or underenhanced and cause color distortion, both of ...
An Image Enhancement Approach Using Retinex and YIQ
ITCS '09: Proceedings of the 2009 International Conference on Information Technology and Computer Science - Volume 01On a basis of Retinex and YIQ color space, a new approach of color image enhancement is presented in this paper. The original image in the RGB color space is converted to the YIQ color space, and the multi-scale Retinex (MSR) algorithm is introduced to ...






Comments