

Metrics

PSNR

def PSNR(y_true, y_pred, MAXp)

Evaluates the PSNR value: PSNR = 20 * log10(MAXp) - 10 * log10(MSE).

Args

• y_true: ground truth.

• y_pred: predicted value.

• MAXp: maximum value of the pixel range (default=1).

RGB_to_Y

def RGB_to_Y(image)

Image has values from 0 to 1.

PSNR_Y

def PSNR_Y(y_true, y_pred, MAXp)

Evaluates the PSNR value on the Y channel: PSNR = 20 * log10(MAXp) - 10 * log10(MSE).

Args

• y_true: ground truth.

• y_pred: predicted value.

• MAXp: maximum value of the pixel range (default=1).