# PairLIE 论文详解

论文为 2023CVPR 的 Learning a Simple Low-light Image Enhancer from Paired Low-light Instances. 论文链接如下:

openaccess.thecvf.com/content/CVPR2023/papers/Fu_Learning_a_Simple_Low-Light_Image_Enhancer_From_Paired_Low-Light_Instances_CVPR_2023_paper.pdf

# 出发点

1.However, collecting high-quality reference maps in real-world scenarios is time-consuming and expensive.

出发点 1:在低光照领域,从现实世界中获取高质量的参考照片进行监督学习,既费时又困难,成本昂贵。

因为获得低光环境的照片是容易的,而此低光照片对应的亮度较大的参考图片是难得的。

2.To tackle the issues of limited information in a single low-light image and the poor adaptability of handcrafted priors, we propose to leverage paired low-light instances to train the LIE network.

Additionally, twice-exposure images provide useful information for solving the LIE task. As a result, our solution can reduce the demand for handcrafted priors and improve the adaptability of the network.

出发点 2:为了解决手动设置的先验的低适应性,减少手动设置先验的需求,同时提升模型对陌生环境的适应性。

# 创新点

The core insight of our approach is to sufficiently exploit priors from paired low-light images.

Those low-light image pairs share the same scene content but different illumination. Mathematically, Retinex decomposition with low-light image pairs can be expressed as:

image

创新点 1:作者利用两张低光图片进行训练,以充分提取低光图片的信息。

instead of directly imposing the Retinex decomposition on original low-light images, we adopt a simple self-supervised mechanism to remove inappropriate features and implement the Retinex decomposition on the optimized image.

创新点 2:作者基于 Retinex 理论,但是并不循旧地直接运用 Retinex 的分解。作者采用一个简单的自监督机制以实现不合理特征的去除(通常是一些噪音)以及更好地实现 Retinex 理论。

# 模型

image

将两张同一场景不同曝光的低光图片送入训练中,图片 I1 与 I2 先经过 P-Net 去除噪音,得到 i1 与 i2,然后利用 L-Net 与 R-Net 分解为照度 L1 与反射 R1(对应有 L2 与 R2)。

在测试,只需要输入一张低光照图片 I,经过 P-Net 的噪音去除,得到 i,然后用 L-Net 与 R-Net 分解为照度和反射,然后对照度 L 进行增强,操作为 g (L),把增强结果与反射 R 进行元素乘法,得到增强后的图片 Enhanced Image。

# 设计及其损失

Note that, this paper does not focus on designing modernistic network structures. L-Net and R-Net are very similar and simple,

1. 模型使用的 L-Net 与 R-Net 十分简单。整体架构只是单纯的卷积神经网络。

Apart from L-Net and R-Net, we introduce P-Net to remove inappropriate features from the original image. Specifically, the structure of the P-Net is identical to the R-Net.

2,P-Net 被设计用于去除不合理特征。

Lp=I1i122L_p = \mid\mid I_1 - i_1 \mid\mid^2_2

Note that the projection loss needs to cooperate with the other constraints to avoid a trivial solution.i,e.,i1 = I1.

3.Projection Loss:最大程度限制去除不合理特征后的 i1 和原始低光图片 I1 的区别。

这个损失需要避免一个特例,即降噪后图片与原图相同,即未降噪。

Lc=R1R222(1)L_c = \mid\mid R_1 - R_2 \mid\mid^2_2\tag{1}

Since sensor noise hidden in dark regions will be amplified when the contrast is improved.

In our method, the sensor noise can be implicitly removed by Eq. 1.

4.Reflection Loss:通常用传感或摄影设备拍摄低光场景照片会携带一定的设备噪音,这个损失最大限度保证两张图片的反射是相同的,减少传感或摄影设备的影响,这是因为图片场景的内容相同。

这个损失是确保反射的一致性。

L_R=RLi2_2+Ri/stopgrad(L)2_2+LL_02_2+L1L\_R = \mid\mid R \circ L - i \mid\mid^2\_2 + \mid\mid R - i / stopgrad(L)\mid\mid^2\_2 + \mid\mid L - L\_0 \mid\mid^2\_2 + \mid\mid \nabla L \mid\mid_1

RLi22\mid\mid R \circ L - i \mid\mid^2_2is applied to ensure a reasonable decomposition.

Ri/stopgrad(L)22\mid\mid R - i / stopgrad(L) \mid\mid^2_2 is to guide the decomposition.

Specifically, the initialized illumination L0 is calculated via the maximum of the R, G, and B channels:L0=maxcR,G,BIc(x).L_0 = \underset{c \in{R, G, B}}{max} I^c(x).

5.Retinex Loss:Retinex 损失是为了限制分解组块 L-Net 和 R-Net 以满足 Retinex 的理论要求。

本文毕

更新于 阅读次数

请我喝[茶]~( ̄▽ ̄)~*

Runhua Deng 微信支付

微信支付

Runhua Deng alipay

alipay

Runhua Deng paypal

paypal