Article
Computer Science, Artificial Intelligence
Yiming Zhu, Cairong Wang, Chenyu Dong, Ke Zhang, Hongyang Gao, Chun Yuan
Summary: This paper presents a method called HfFlow for solving the problem of lost high-frequency details during image rescaling. By learning the distribution of high-frequency signals and compensating with conditional flow, HfFlow is able to restore rich high-frequency details and outperforms other methods in various evaluation metrics.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Chemistry, Multidisciplinary
Xinxin Pan, Changli Li, Zhigeng Pan, Jingwen Yan, Shiqiang Tang, Xinghui Yin
Summary: This study proposes a retinex-based method for low-light image enhancement, which improves image brightness and contrast by enhancing the illumination map. Experimental results demonstrate that it outperforms other state-of-the-art methods in both subjective vision and quantitative evaluation.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Ping Wang, Zhiwen Wang, Dong Lv, Chanlong Zhang, Yuhang Wang
Summary: The low-illuminance color image enhancement algorithm proposed in this study combines Gabor filter and Retinex theory to enhance images in two different ways before merging them to reduce halo and excessive enhancement, showing promise for enhancing low-light color images.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Zhijie Tang, Lizhou Jiang, Zhihang Luo
Summary: The paper proposes an underwater image enhancement algorithm, which achieves good results in color processing by improving the algorithm, and also improves the contrast and clarity of the image.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Mingqing Xiao, Shuxin Zheng, Chang Liu, Zhouchen Lin, Tie-Yan Liu
Summary: This article introduces a reversible framework for handling image rescaling, which addresses the issue of information loss caused by non-injective downscaling mapping. The framework models the bidirectional degradation and restoration and uses invertible models to generate valid degraded images. The experimental results demonstrate the superior performance of the proposed model in terms of quantitative and qualitative evaluations, as well as image compression.
INTERNATIONAL JOURNAL OF COMPUTER VISION
(2023)
Article
Automation & Control Systems
Dayi Li, Jingchun Zhou, Shiyin Wang, Dehuan Zhang, Weishi Zhang, Raghad Alwadai, Fayadh Alenezi, Prayag Tiwari, Taian Shi
Summary: This paper develops an adaptive weighted multiscale retinex (AWMR) method to enhance underwater images, improving both the quality and performance compared to current state-of-the-art techniques.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Hayk A. Gasparyan, Sargis A. Hovhannisyan, Stepan V. Babayan, Sos S. Agaian
Summary: Images captured in low-light conditions often suffer from issues such as low brightness, low signal-to-noise ratio, and color distortion. This paper proposes a new framework called RSD-Net, which includes innovative algorithms to enhance the visibility and quality of low-light images. The proposed framework achieves significant improvement compared to other methods and enhances face detection accuracy in low-quality environments.
Article
Multidisciplinary Sciences
Xiwen Liang, Xiaoyan Chen, Keying Ren, Xia Miao, Zhihui Chen, Yutao Jin
Summary: In this work, the authors propose a novel Adaptive Frequency Decomposition Network (AFDNet) for enhancing low-light images. They utilize an Adaptive Frequency Decomposition (AFD) module to extract low and high frequency information and improve contrast, noise, and detail. Extensive experiments show that AFDNet outperforms existing state-of-the-art methods both quantitatively and visually.
SCIENTIFIC REPORTS
(2023)
Article
Computer Science, Artificial Intelligence
Yuetian Shi, Bin Fu, Nan Wang, Yaxiong Chen, Jie Fang
Summary: This study proposes a novel multispectral image restoration algorithm that considers the impact of meteorological factors on image quality, resulting in improved performance and robustness compared to other algorithms. Experimental results demonstrate the effectiveness of the algorithm on real data.
COGNITIVE COMPUTATION
(2023)
Article
Computer Science, Information Systems
Yi Han, Xiangyong Chen, Yi Zhong, Yanqing Huang, Zhuo Li, Ping Han, Qing Li, Zhenhui Yuan
Summary: This paper proposes a histogram equalization-multiscale Retinex combination approach (HE-MSR-COM) to solve the blur edge problem of HE and the uncertainty in selecting parameters for image illumination enhancement in MSR. Experimental results show that HE-MSR-COM improves image quality by 23.95% and 10.6% in two datasets, respectively, compared with HE, CLAHE, MSR, and GC.
Article
Multidisciplinary Sciences
Yazhong Si, Fan Yang, Zhao Liu
Summary: This paper proposes a novel image enhancement algorithm based on fusion strategy to address the issues of poor contrast and color distortion in outdoor images captured in sand dust weather. The algorithm includes sand removal using an improved color correction algorithm and dust elimination using a residual-based convolutional neural network. Theoretical analysis and experimental results demonstrate that the proposed fusion strategy effectively corrects the overall yellowing hue and removes dust haze disturbance.
SCIENTIFIC REPORTS
(2022)
Article
Chemistry, Analytical
Hongjun Tan, Dongxiu Ou, Lei Zhang, Guochen Shen, Xinghua Li, Yuqing Ji
Summary: This paper proposes an enhancement processing method based on temperature and pixel features of infrared images. It includes threshold denoising, salient components enhancement, and utilizing salient features of targets to improve image contrast and noise distribution. Experimental results demonstrate the high accuracy and adaptability of the proposed method in target detection and recognition.
Article
Computer Science, Interdisciplinary Applications
Lei Chen, Chen Tang, Min Xu, Zhenkun Lei
Summary: The paper proposes an enhancement and denoising strategy for low-quality medical images based on the sequence decomposition Retinex model and the inverse haze removal approach, improving visibility by effectively enhancing contrast and suppressing noise. Experimental results demonstrate a splendid balance between enhancement and denoising, with performance superior to that of several state-of-the-art methods.
MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING
(2021)
Article
Engineering, Electrical & Electronic
Xiaochun Lei, Weiliang Mai, Junlin Xie, He Liu, Zetao Jiang, Zhaoting Gong, Chang Lu, Linjun Lu
Summary: This paper proposes an algorithm for low illumination enhancement, which uses light curve estimation and image fusion techniques to improve image brightness and restore details, while applying a total variation loss function to eliminate noise. Experimental results demonstrate the competitive performance of the method on multiple datasets.
SIGNAL IMAGE AND VIDEO PROCESSING
(2023)
Article
Engineering, Civil
Xianjun Hu, Jing Wang, Guilian Li
Summary: With the rapid development of AI and big traffic data, data-driven intelligent maritime transportation has gained attention in both industry and academia. In this work, a contrastive learning framework is proposed for enhancing visibility under hazy imaging conditions in maritime transportation systems, improving visual quality and ship detection accuracy.
JOURNAL OF ADVANCED TRANSPORTATION
(2022)