Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement
The goal of low-light image enhancement (LLIE) is to enhance perception to restore normal-light images. The primary emphasis of earlier LLIE methods was on enhancing the illumination while paying less attention to the color distortions and noise in the dark. In comparison to the ground truth, the re...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/11/19/4194 |
| _version_ | 1797575540642676736 |
|---|---|
| author | Zixuan Sun Shenglong Hu Huihui Song Peng Liang |
| author_facet | Zixuan Sun Shenglong Hu Huihui Song Peng Liang |
| author_sort | Zixuan Sun |
| collection | DOAJ |
| description | The goal of low-light image enhancement (LLIE) is to enhance perception to restore normal-light images. The primary emphasis of earlier LLIE methods was on enhancing the illumination while paying less attention to the color distortions and noise in the dark. In comparison to the ground truth, the restored images frequently exhibit inconsistent color and residual noise. To this end, this paper introduces a Wasserstein contrastive regularization method (WCR) for LLIE. The WCR regularizes the color histogram (CH) representation of the restored image to keep its color consistency while removing noise. Specifically, the WCR contains two novel designs including a differentiable CH module (DCHM) and a WCR loss. The DCHM serves as a modular component that can be easily integrated into the network to enable end-to-end learning of the image CH. Afterwards, to ensure color consistency, we utilize the Wasserstein distance (WD) to quantify the resemblance of the learnable CHs between the restored image and the normal-light image. Then, the regularized WD is used to construct the WCR loss, which is a triplet loss and takes the normal-light images as positive samples, the low-light images as negative samples, and the restored images as anchor samples. The WCR loss pulls the anchor samples closer to the positive samples and simultaneously pushes them away from the negative samples so as to help the anchors remove the noise in the dark. Notably, the proposed WCR method was only used for training, and was shown to achieve high performance and high speed inference using lightweight networks. Therefore, it is valuable for real-time applications such as night automatic driving and night reversing image enhancement. Extensive evaluations on benchmark datasets such as LOL, FiveK, and UIEB showed that the proposed WCR method achieves superior performance, outperforming existing state-of-the-art methods. |
| first_indexed | 2024-03-10T21:40:04Z |
| format | Article |
| id | doaj.art-ac50a8239822435082596324460cbd0d |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-10T21:40:04Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-ac50a8239822435082596324460cbd0d2023-11-19T14:44:38ZengMDPI AGMathematics2227-73902023-10-011119419410.3390/math11194194Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image EnhancementZixuan Sun0Shenglong Hu1Huihui Song2Peng Liang3B-DAT and CICAEET, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaB-DAT and CICAEET, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaB-DAT and CICAEET, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaCollege of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510000, ChinaThe goal of low-light image enhancement (LLIE) is to enhance perception to restore normal-light images. The primary emphasis of earlier LLIE methods was on enhancing the illumination while paying less attention to the color distortions and noise in the dark. In comparison to the ground truth, the restored images frequently exhibit inconsistent color and residual noise. To this end, this paper introduces a Wasserstein contrastive regularization method (WCR) for LLIE. The WCR regularizes the color histogram (CH) representation of the restored image to keep its color consistency while removing noise. Specifically, the WCR contains two novel designs including a differentiable CH module (DCHM) and a WCR loss. The DCHM serves as a modular component that can be easily integrated into the network to enable end-to-end learning of the image CH. Afterwards, to ensure color consistency, we utilize the Wasserstein distance (WD) to quantify the resemblance of the learnable CHs between the restored image and the normal-light image. Then, the regularized WD is used to construct the WCR loss, which is a triplet loss and takes the normal-light images as positive samples, the low-light images as negative samples, and the restored images as anchor samples. The WCR loss pulls the anchor samples closer to the positive samples and simultaneously pushes them away from the negative samples so as to help the anchors remove the noise in the dark. Notably, the proposed WCR method was only used for training, and was shown to achieve high performance and high speed inference using lightweight networks. Therefore, it is valuable for real-time applications such as night automatic driving and night reversing image enhancement. Extensive evaluations on benchmark datasets such as LOL, FiveK, and UIEB showed that the proposed WCR method achieves superior performance, outperforming existing state-of-the-art methods.https://www.mdpi.com/2227-7390/11/19/4194low-light image enhancementWasserstein contrastive regularizationdifferentiable color histogram module |
| spellingShingle | Zixuan Sun Shenglong Hu Huihui Song Peng Liang Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement Mathematics low-light image enhancement Wasserstein contrastive regularization differentiable color histogram module |
| title | Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement |
| title_full | Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement |
| title_fullStr | Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement |
| title_full_unstemmed | Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement |
| title_short | Learning Wasserstein Contrastive Color Histogram Representation for Low-Light Image Enhancement |
| title_sort | learning wasserstein contrastive color histogram representation for low light image enhancement |
| topic | low-light image enhancement Wasserstein contrastive regularization differentiable color histogram module |
| url | https://www.mdpi.com/2227-7390/11/19/4194 |
| work_keys_str_mv | AT zixuansun learningwassersteincontrastivecolorhistogramrepresentationforlowlightimageenhancement AT shenglonghu learningwassersteincontrastivecolorhistogramrepresentationforlowlightimageenhancement AT huihuisong learningwassersteincontrastivecolorhistogramrepresentationforlowlightimageenhancement AT pengliang learningwassersteincontrastivecolorhistogramrepresentationforlowlightimageenhancement |