Implementation:Kornia Kornia Rgb To Grayscale

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

Overview

Concrete tool for differentiable RGB-to-grayscale conversion provided by Kornia's color module.

Description

The rgb_to_grayscale function converts RGB tensors to single-channel grayscale using configurable weights. Default weights follow ITU-R BT.601 (0.299, 0.587, 0.114). The operation is differentiable and supports batched inputs. It accepts (*, 3, H, W) tensors and outputs (*, 1, H, W).

Usage

Import when preprocessing images for feature detection, edge detection, or any grayscale-requiring operation.

Code Reference

Source Location

Repository: kornia
File: kornia/color/gray.py
Lines: L58-109

Signature

def rgb_to_grayscale(
    image: torch.Tensor,
    rgb_weights: Optional[torch.Tensor] = None
) -> torch.Tensor

Import

from kornia.color import rgb_to_grayscale

I/O Contract

Inputs

Name	Type	Required	Description
image	torch.Tensor	Yes	RGB tensor of shape (*, 3, H, W)
rgb_weights	Optional[torch.Tensor]	No	Custom channel weights for the grayscale conversion

Outputs

Name	Type	Description
output	torch.Tensor	Grayscale image of shape (*, 1, H, W)

Usage Examples

Basic Conversion

import torch
from kornia.color import rgb_to_grayscale

rgb_image = torch.rand(1, 3, 256, 256)
gray_image = rgb_to_grayscale(rgb_image)
# gray_image.shape == (1, 1, 256, 256)

Custom Weights

import torch
from kornia.color import rgb_to_grayscale

rgb_image = torch.rand(1, 3, 256, 256)
weights = torch.tensor([0.333, 0.333, 0.334])
gray_image = rgb_to_grayscale(rgb_image, rgb_weights=weights)

Batch Conversion

import torch
from kornia.color import rgb_to_grayscale

batch = torch.rand(8, 3, 256, 256)
gray_batch = rgb_to_grayscale(batch)
# gray_batch.shape == (8, 1, 256, 256)

Related Pages

Implements Principle

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment