Instance Normalization

Introduced by Ulyanov et al. in Instance Normalization: The Missing Ingredient for Fast Stylization

Instance Normalization (also known as contrast normalization) is a normalization layer where:

$$ y_{tijk} = \frac{x_{tijk} - \mu_{ti}}{\sqrt{\sigma_{ti}^2 + \epsilon}}, \quad \mu_{ti} = \frac{1}{HW}\sum_{l=1}^W \sum_{m=1}^H x_{tilm}, \quad \sigma_{ti}^2 = \frac{1}{HW}\sum_{l=1}^W \sum_{m=1}^H (x_{tilm} - \mu_{ti})^2. $$

This prevents instance-specific mean and covariance shift simplifying the learning process. Intuitively, the normalization process allows to remove instance-specific contrast information from the content image in a task like image stylization, which simplifies generation.

Source: Instance Normalization: The Missing Ingredient for Fast Stylization

Read Paper See Code

Papers

Paper	Code	Results	Date	Stars

Tasks

Task	Papers	Share
Image-to-Image Translation	72	11.32%
Domain Adaptation	41	6.45%
Style Transfer	39	6.13%
Image Generation	38	5.97%
Semantic Segmentation	33	5.19%
Unsupervised Domain Adaptation	18	2.83%
Object Detection	15	2.36%
Voice Conversion	14	2.20%
Disentanglement	12	1.89%

Usage Over Time

This feature is experimental; we are continuously improving our matching algorithm.

Components

Component	Type	Add Remove
🤖 No Components Found	You can add them if they exist; e.g. Mask R-CNN uses RoIAlign

Categories

Add Remove

Normalization