Fork me on GitHub

Namespace: NRMSE

NRMSE

The NRMSE (normalized root mean square error) - reference-based image metric.
Source:

Members

(static) defaultTolerance

The allowed size difference when comparing images. The leftover extra pixels of the larger image will be ignored.
Source:

(static) pixelMaxValue

The maximum value of a pixel component. 255 for 8 bit depth
Source:

Methods

(static) compare(image1, image2, tolerance)

Compares two images.
Parameters:
Name Type Description
image1 * Preferably the ground truth / reference (convention, shouldnt matter otherwise)
image2 * An image to compare
tolerance * The allowed size ratio difference, default is NRMSE.defaultTolerance.
Source:
Returns:
An object with the calculate metric values of MSE, PSNR, RMSE, NRMSE, Inverted NRMSE, NMSE, and Inverted NMSE.

(static) log10()

Utility function that performs a Log base 10 calculation.
Source:

(static) psnr(mse, max)

Calculates PSNR (Peak Signal-to-Noise Ratio).
Parameters:
Name Type Description
mse * a Mean Squared Error (MSE) value
max * the maximum value of a pixel component, default is NRMSE.pixelMaxValue
Source:
Returns:
the calculated value

↑ Top