Implementation:Kornia Kornia Luv Conversion

Knowledge Sources	Kornia
Domains	Vision, Color_Processing
Last Updated	2026-02-09 15:00 GMT

Overview

This module provides differentiable conversion between RGB and CIE L*u*v* color space using the D65 illuminant.

Description

luv.py is a module in the Kornia library's color subpackage implementing bidirectional RGB-to-Luv (CIE L*u*v*) color space conversion, translated from scikit-image. The conversion pipeline follows: sRGB -> linear RGB -> XYZ -> Luv, using the D65 white point reference. The rgb_to_luv function computes L (lightness), u, and v chromaticity coordinates with an epsilon parameter for numerical stability. The luv_to_rgb function reverses the pipeline: Luv -> XYZ -> linear RGB -> sRGB. Each function has a corresponding nn.Module wrapper (RgbToLuv and LuvToRgb). The module depends on rgb.py and xyz.py for intermediate conversions.

Usage

Import this module when you need perceptually uniform chromaticity coordinates, for example in color constancy algorithms, illumination estimation, or any application where uniform perceptual distances in both lightness and chromaticity are important.

Code Reference

Source Location

• Repository: Kornia
• File: kornia/color/luv.py
• Lines: 1-204

Signature

def rgb_to_luv(image: torch.Tensor, eps: float = 1e-12) -> torch.Tensor: ...
def luv_to_rgb(image: torch.Tensor, eps: float = 1e-12) -> torch.Tensor: ...

class RgbToLuv(nn.Module):
    def forward(self, image: torch.Tensor) -> torch.Tensor: ...

class LuvToRgb(nn.Module):
    def forward(self, image: torch.Tensor) -> torch.Tensor: ...

Import

from kornia.color import rgb_to_luv, luv_to_rgb
from kornia.color import RgbToLuv, LuvToRgb

I/O Contract

Inputs

Name	Type	Required	Description
image (rgb_to_luv)	torch.Tensor	Yes	RGB image with shape (*, 3, H, W). Values in range [0, 1].
eps (rgb_to_luv)	float	No	Epsilon for numerical stability. Default: 1e-12.
image (luv_to_rgb)	torch.Tensor	Yes	Luv image with shape (*, 3, H, W).
eps (luv_to_rgb)	float	No	Epsilon for numerical stability. Default: 1e-12.

Outputs

Name	Type	Description
rgb_to_luv return	torch.Tensor	Luv image with shape (*, 3, H, W). L (lightness), u, and v chromaticity.
luv_to_rgb return	torch.Tensor	sRGB image with shape (*, 3, H, W).

Usage Examples

Basic Usage

import torch
from kornia.color import rgb_to_luv, luv_to_rgb

# Convert RGB to Luv
rgb = torch.rand(1, 3, 128, 128)
luv = rgb_to_luv(rgb)
print(luv.shape)  # torch.Size([1, 3, 128, 128])

# Convert Luv back to RGB
rgb_back = luv_to_rgb(luv)
print(rgb_back.shape)  # torch.Size([1, 3, 128, 128])

# Using nn.Module wrappers
from kornia.color import RgbToLuv, LuvToRgb
to_luv = RgbToLuv()
to_rgb = LuvToRgb()
luv_out = to_luv(rgb)
rgb_out = to_rgb(luv_out)

Related Pages

• Principle:Kornia_Kornia_Luv_Color_Space