当前位置: X-MOL 学术Signal Image Video Process. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Frequency-domain loss function for deep exposure correction of dark images
Signal, Image and Video Processing ( IF 2.3 ) Pub Date : 2021-05-15 , DOI: 10.1007/s11760-021-01915-4
Ojasvi Yadav 1 , Koustav Ghosal 1 , Sebastian Lutz 1 , Aljosa Smolic 1
Affiliation  

We address the problem of exposure correction of dark, blurry and noisy images captured in low-light conditions in the wild. Classical image-denoising filters work well in the frequency space but are constrained by several factors such as the correct choice of thresholds and frequency estimates. On the other hand, traditional deep networks are trained end to end in the RGB space by formulating this task as an image translation problem. However, that is done without any explicit constraints on the inherent noise of the dark images and thus produces noisy and blurry outputs. To this end, we propose a DCT/FFT-based multi-scale loss function, which when combined with traditional losses, trains a network to translate the important features for visually pleasing output. Our loss function is end to end differentiable, scale-agnostic and generic; i.e., it can be applied to both RAW and JPEG images in most existing frameworks without additional overhead. Using this loss function, we report significant improvements over the state of the art using quantitative metrics and subjective tests.



中文翻译:

暗图像深度曝光校正的频域损失函数

我们解决了在野外低光照条件下拍摄的黑暗、模糊和嘈杂图像的曝光校正问题。经典的图像去噪滤波器在频率空间中运行良好,但受到多个因素的限制,例如正确选择阈值和频率估计。另一方面,传统的深度网络在 RGB 空间中通过将此任务表述为图像翻译问题进行端到端的训练。但是,这是在没有任何明确的情况下完成的限制暗图像的固有噪声,从而产生噪声和模糊的输出。为此,我们提出了一种基于 DCT/FFT 的多尺度损失函数,当与传统损失相结合时,可以训练网络转换重要特征以获得视觉上令人愉悦的输出。我们的损失函数是端到端可微的、与规模无关的和通用的;即,它可以应用于大多数现有框架中的 RAW 和 JPEG 图像,而无需额外开销。使用此损失函数,我们使用定量指标和主观测试报告了对现有技术的显着改进。

更新日期:2021-05-15
down
wechat
bug