【论文笔记】LIME: low-light image enhancement via illumination map estimation

Traditional Methods & Unsolved Problems

• Histogram equalization (HE)
  • focus on contrast enhancement instead of exploiting real illumination causes, having the risk of over- and under- enhancement
• Gamma correction
  • carried out on each pixel individually
• single-scale Retinex (SSR) and multi-scale Retinex (MSR)
  • looks unnatural and frequently appears to be over-enhanced
• adjust the illumination by fusing multiple derivations of the initially estimated illumination map
  • lose the realism of regions with rich textures
• dehaze inverted low-light image
  • lacking in physical explanation

Solutions

• Retinex-based methods but only estimates one factor, say the illumination
  • shrink the solution space and reduce the computational cost
• exploit the structure of illumination to refine illumination map

Retinex

$$\mathbf{L}=\mathbf{R}\odot \mathbf{T}$$
where

• $\mathbf{L}$ are the captured low-light image
• $\mathbf{R}$ are the desired recovery (aka reflectance)
• $\mathbf{T}$ represents the illumination map
• $\odot$ means element-wise multiplication

Assume: for color images, three channels share the same illumination map. $\mathbf{T}$ and $\hat{\mathbf{T}}$ to represent one-channel and three-channel illumination maps.

Illumination Estimation

T ^ ( x ) ← max ⁡ c ∈ { R , G , B } L c ( x ) (1) \hat {\mathbf T}(x) \leftarrow \max_{c∈\{R,G,B\}} \mathbf L ^c(x)\tag 1 T^(x)←c∈{ R,G,B}max​Lc(x)(1)
so $$\mathbf{R}(x)=\frac{\mathbf{L}(x)}{{\max }_c{\mathbf{L}}^c(x)+ϵ}\text{\hspace{1em}(2)}$$
In dehaze model:
$$1-\mathbf{L}=(1-\mathbf{R})\odot \tilde{\mathbf{T}}+a(1-\tilde{\mathbf{T}})\text{\hspace{1em}(3)}$$
where

• $1-\mathbf{L}$ is inverted low-light image (similar to haze image)

with dark channel prior
$$\tilde{\mathbf{T}}(x)\leftarrow 1-\underset{c}{\min }\frac{1-{\mathbf{L}}^c(x)}{a}=1-\frac{1}{a}+\underset{c}{\max }$$