低光图像通过随机伽马调整和泊松噪声添加生成,视频则从e-Lab Video Data Set中裁剪得到。 实验设置:在Titan-X GPU上进行训练,使用ADAM优化器,学习率为0.002,批量大小为24,输入图像值缩放到[0,1]。使用SSIM和MS-SSIM作为结构损失的计算基础,选择VGG-19网络的第四卷积层的输出作为上下文损失的提取层。 结果与分析 低光
论文阅读笔记---MBLLEN: Low-light Image/Video Enhancement Using CNNs,程序员大本营,技术文章内容聚合第一站。
论文阅读笔记---MBLLEN: Low-light Image/Video Enhancement Using CNNs 摘要:提出了一种基于深度学习的微光图像增强方法。由于难以同时处理包括亮度、对比度、伪影和噪声在内的各种因素,这个问题具有挑战性。为了解决这一问题,我们提出了多分支微光增强网络(MBLLEN)。其核心思想是提取不同层次的丰富特征,通过多个子网络...
In recent years, there has been a shift towards deep learning-based approaches in the research on low-light image enhancement. LLNet25is a pioneering work by LLIE that performs contrast enhancement and denoising based on a depth autoencoder. However, the relationship between real-world illumination...
Code for “MBLLEN: Low-light Image/Video Enhancement Using CNNs”, BMVC 2018. - Lvfeifan/MBLLEN
Image and video enhancementMobile devicesSuper-resolutionTop-hat transformImage and video enhancement under low-light conditions is challenging, as the task involves more than just brightness adjustment. Without addressing issues such as artifacts, distortions, and noise in dark regions, brightness ...
IEEE Transactions on image processing (2000) J. Park, J.-Y. Lee, D. Yoo, I. So Kweon, Distort-and-recover: Color enhancement using deep reinforcement learning, in:... F. Lv, F. Lu, J. Wu, C. Lim, Mbllen: Low-light image/video enhancement using cnns., in: BMVC, 2018, p.....
Lv, F., Lu, F., Wu, J., Lim, C.: MBLLEN: Low-light image/video enhancement using CNNs. In: British Machine Vision Conference (BMVC) (2018) Google Scholar Maggioni, M., Boracchi, G., Foi, A., Egiazarian, K.O.: Video denoising, deblocking, and enhancement through separable 4...
Mbllen: Low-light image/video enhancement using cnns. In BMVC, pages 1–13, 2018. 6 [21] Kede Ma, Kai Zeng, and Zhou Wang. Perceptual qual- ity assessment for multi-exposure image fusion. IEEE TIP, 24(11):3345–3356, 2015. 6 [22] Long Ma, Te...
论文阅读笔记---MBLLEN: Low-light Image/Video Enhancement Using CNNs 摘要:提出了一种基于深度学习的微光图像增强方法。由于难以同时处理包括亮度、对比度、伪影和噪声在内的各种因素,这个问题具有挑战性。为了解决这一问题,我们提出了多分支微光增强网络(MBLLEN)。其核心思想是提取不同层次的丰富特征,通过多个子网络...