LVE-S2D: Low-Light Video Enhancement From Static to Dynamic

Peng, Bo; Zhang, Xuanyu; Lei, Jianjun; Zhang, Zhe; Ling, Nam; Huang, Qingming

doi:10.1109/tcsvt.2022.3190916

Cited by 28 publications

(5 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, deep neural networks have been widely used in various fields, such as 3D shape clustering [32], quality assessment [33], and video enhancement [34]. Taking advantage of deep neural networks, researchers have conducted preliminary explorations of different visual tasks in subterranean scenes, which mainly contain the two categories of simultaneous localization and mapping (SLAM) [35][36][37] and object detection [38][39][40].…”

Section: Visual Tasks In Subterranean Scenesmentioning

confidence: 99%

Domain Adaptive Subterranean 3D Pedestrian Detection via Instance Transfer and Confidence Guidance

Liu,

Zheng,

Qin

et al. 2024

Electronics

View full text Add to dashboard Cite

With the exploration of subterranean scenes, determining how to ensure the safety of subterranean pedestrians has gradually become a hot research topic. Considering the poor illumination and lack of annotated data in subterranean scenes, it is essential to explore the LiDAR-based domain adaptive detectors for localizing the spatial location of pedestrians, thus providing instruction for evacuation and rescue. In this paper, a novel domain adaptive subterranean 3D pedestrian detection method is proposed to adapt pre-trained detectors from the annotated road scenes to the unannotated subterranean scenes. Specifically, an instance transfer-based scene updating strategy is designed to update the subterranean scenes by transferring instances from the road scenes to the subterranean scenes, aiming to create sufficient high-quality pseudo labels for fine-tuning the pre-trained detector. In addition, a pseudo label confidence-guided learning mechanism is constructed to fully utilize pseudo labels of different qualities under the guidance of confidence scores. Extensive experiments validate the superiority of our proposed domain adaptive subterranean 3D pedestrian detection method.

show abstract

Section: Visual Tasks In Subterranean Scenesmentioning

confidence: 99%

Domain Adaptive Subterranean 3D Pedestrian Detection via Instance Transfer and Confidence Guidance

Liu,

Zheng,

Qin

et al. 2024

Electronics

View full text Add to dashboard Cite

show abstract

“…Recently, deep learning technology has developed rapidly and achieved impressive success in various tasks, such as quality enhancement [9,10], content generation [11,12], and object detection [13,14]. Inspired by this, researchers have explored many deep learningbased video coding methods [15][16][17].…”

Section: Introductionmentioning

confidence: 99%

End-to-End Light Field Image Compression with Multi-Domain Feature Learning

Ye,

Li,

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

Recently, end-to-end light field image compression methods have been explored to improve compression efficiency. However, these methods have difficulty in efficiently utilizing multi-domain features and their correlation, resulting in limited improvement in compression performance. To address this problem, a novel multi-domain feature learning-based light field image compression network (MFLFIC-Net) is proposed to improve compression efficiency. Specifically, an EPI-based angle completion module (E-ACM) is developed to obtain a complete angle feature by fully exploring the angle information with a large disparity contained in the epipolar plane image (EPI) domain. Furthermore, in order to effectively reduce redundant information in the light field image, a spatial-angle joint transform module (SAJTM) is proposed to reduce redundancy by modeling the intrinsic correlation between spatial and angle features. Experimental results demonstrate that MFLFIC-Net achieves superior performance on MS-SSIM and PSNR metrics compared to public state-of-the-art methods.

show abstract

“…The aims of low-light image enhancement are to enhance the visibility and contrast of low-light images [13]. Numerous conventional methods for enhancing low-light images have been implemented [14,15].…”

Section: Introductionmentioning

confidence: 99%

“…Conventional processing methods typically depend on specific statistical models and assumptions, such as the Retinex model [15] and histogram equalization [16]. Nevertheless, these techniques necessitate parameter adjustments tailored to varying application scenarios, and the imprecise statistical models can easily introduce excessive artifacts [13]. Traditional methods work well for low-contrast images (such as overall darker images); however, they often do not take into account the effects of spatially varing illumination, leading to an unbalanced intensity distribution in local areas [17].…”

Section: Introductionmentioning

confidence: 99%

LLE-NET: A Low-Light Image Enhancement Algorithm Based on Curve Estimation

Cao,

2024

Mathematics

View full text Add to dashboard Cite

Low-light image enhancement is very significant for vision tasks. We introduce Low-light Image Enhancement via Deep Learning Network (LLE-NET), which employs a deep network to estimate curve parameters. Cubic curves and gamma correction are employed for enhancing low-light images. Our research trains a lightweight network to estimate the parameters that determine the correction curve. By the results of the deep learning network, accurate correction curves are confirmed, which are used for the per-pixel correction of RGB channels. The image enhanced by our models closely resembles the input image. To further accelerate the inferring speed of the low-light enhancement model, a low-light enhancement model based on gamma correction is proposed with one iteration. LLE-NET exhibits remarkable inference speed, achieving 400 fps on a single GPU for images sized 640×480×3 while maintaining pleasing enhancement quality. The enhancement model based on gamma correction attains an impressive inference speed of 800 fps for images sized 640×480×3 on a single GPU.

show abstract

LVE-S2D: Low-Light Video Enhancement From Static to Dynamic

Cited by 28 publications

References 40 publications

Domain Adaptive Subterranean 3D Pedestrian Detection via Instance Transfer and Confidence Guidance

Domain Adaptive Subterranean 3D Pedestrian Detection via Instance Transfer and Confidence Guidance

End-to-End Light Field Image Compression with Multi-Domain Feature Learning

LLE-NET: A Low-Light Image Enhancement Algorithm Based on Curve Estimation

Contact Info

Product

Resources

About