Image Normalization

Image

Image

None

None

Input image

Normalized image

Image

None

None

Normalized image.

Normalization method

String

['Min-Max Normalization', 'Mean-Variance Normalization', 'Max Absolute Value Normalization', 'L1 Norm Normalization', 'L2 Norm Normalization', 'Logarithmic Normalization']

Min-Max Normalization

Setting the normalization method, maximum minimum normalization is the most common method, image = (image - minimum)/(maximum - minimum) * norm_max, norm_max is set to 1 or 255, mean variance normalization, image = (image - average)/variance * norm_max, norm_max is set to 1 or 255, maximum absolute value normalization, image divided by (Chebyshev distance) L ∞ norm (maximum absolute value) and multiplied by norm_max, norm_max is usually set to 1 or 255, L1 norm normalized, the image is divided by the normalized array (Manhattan distance) L1-norm (sum of absolute values), norm_max can be set to a large number, L2 norm normalized, the image is divided by the normalized array (Euclidean distance) L2-norm, norm_max can be set to a large number log normalized, the image is taken as the logarithm of 10 and then divided by the maximum value is taken as the logarithm of 10, then multiplied by norm_max, norm_max can be set to a large number.

Normalized range max

Float

[1, 10000000]

255

Set the maximum value of the normalized range.

Enable truncation

Bool

[True, False]

False

After enabling, the normalization result will be truncated.

Truncated minimum

Float

[-100000, 100000]

0

Set the truncation minimum value, if the normalized result value is less than this value, set the value to this value.

Truncated maximum

Float

[-100000, 100000]

255

Set the truncation maximum value, if the normalized result value is greater than this value, set the value to this value.

Enable node

Bool

None

True

Turn on node functionality.