Deep Retinex Decomposition for Low-Light Enhancement(Retinex-Net)的损失函数
-
-
- S t r u c t u r e Structure Structure- A w a r e S m o o t h n e s s L o s s Aware\ Smoothness \ Loss Aware Smoothness Loss(结构平滑损失)
- I n v a r i a b l e R e f l e c t a n c e l o s s Invariable \ Reflectance\ loss Invariable Reflectance loss (反射不变损失)
- R e c o n s t r u c t i o n L o s s Reconstruction\ Loss Reconstruction Loss(重构损失)
-
经典的Retinex理论为人类的颜色感知建模。这个理论假设观测到的图像可以分解为反射和照度两部分,那么它可以表示为 S = R ο I S =R \omicron I S=RοI
其中,S代表原图像,R代表反射,I代表光照。o代表逐项素乘积。文章中的损失函数包含三部分, R e c o n s t r u c t i o n L o s s Reconstruction\ Loss Reconstruction Loss L r e c o n L_{recon} Lrecon , I n v a r i a b l e R e f l e c t a n c e L o s s Invariable \ Reflectance\ Loss Invariable Reflectance Loss L i r L_{ir} Lir, 和 L l l u m i n a t i o n S m o o t h n e s s L o s s Lllumination\ Smoothness\ Loss Llluminat