Dynamic Mode

BrightnessContrast Example

This example demonstrates usage of brightness_contrast operator.

Brightness and Contrast Adjustment

Brightness and contrast can be adjusted using three parameters:

• contrast - makes the dark parts darker and bright parts brighter while keeping the gray (at half dynamic range of the input type) constant

• brightness - multiplies the intensities

• brightness_shift - shifts the pixel intensities

The formula is given as:

\[Out = brightness\_shift \cdot {Max}_{out} + brightness \cdot ({contrast\_center} + {contrast} \cdot ({in} - {contrast\_center}))\]

where \({contrast\_center}\) denotes a value that is unaffected by contrast changes; by default, it’s at half of the input’s dynamic range (e.g. 128 for uint8) and \({Max}_{out}\) is the maximum positive value of the output type (for integer types) or 1 (for floating point) - but it can also be specified as an argument for the operator.

The operator can be used to adjust contrast and brightness as well as to compute negative images (brightness = -1, brightness_shift = 1).

from pathlib import Path

import matplotlib.pyplot as plt
import nvidia.dali.experimental.dynamic as ndd

image_dir = Path("../../data/images")
image_path = next(image_dir.glob("**/*.jpg"))

def display(original: ndd.Tensor, converted: ndd.Tensor):
    fig, ax = plt.subplots(1, 2, figsize=(12, 6))
    ax[0].imshow(original.cpu())
    ax[0].set_title("Original")
    ax[0].axis("off")
    ax[1].imshow(converted.cpu())
    ax[1].set_title("Adjusted")
    ax[1].axis("off")

GPU Example: Color Inversion

Let’s apply brightness/contrast adjustment on the GPU to create a color inversion effect. Setting brightness=-1 and brightness_shift=1 inverts the colors, while contrast=1.5 increases the contrast.

image = ndd.imread(image_path, device="gpu")
converted = ndd.brightness_contrast(
    image,
    contrast=1.5,
    brightness=-1,
    brightness_shift=1,
)
display(image, converted)

CPU Example: Contrast and Brightness Adjustment

Now let’s run on the CPU with different parameters. Here we reduce contrast (contrast=0.4) and shift brightness up (brightness_shift=0.3), with a custom contrast center point.

image_cpu = ndd.imread(image_path, device="cpu")
converted_cpu = ndd.brightness_contrast(
    image_cpu,
    brightness_shift=0.3,
    contrast=0.4,
    contrast_center=100,
)
display(image_cpu, converted_cpu)