SegmentationVisualizationConfig

Options & Configuration · AiDotNet.ComputerVision.Segmentation.Common

Visualization settings and utilities for segmentation outputs.

For Beginners

After segmenting an image, you usually want to visualize the result. This class provides configuration for creating color-coded overlay images where each class or instance gets a distinct color. The overlay can be blended with the original image to see both the segmentation and the original content.

Constructors

Constructor Summary
SegmentationVisualizationConfig()

Properties

Property Summary
ValueTuple<byte, byte, byte> BackgroundColor { get; set; } Background color for areas with no segmentation (R, G, B).
bool DrawContours { get; set; } Whether to draw contours/boundaries around segmented regions.
bool ShowBoundingBoxes { get; set; } Whether to display instance bounding boxes.
bool ShowLabels { get; set; } Whether to display class labels on the visualization.
bool ShowScores { get; set; } Whether to display confidence scores alongside labels.
bool UseFixedPalette { get; set; } Whether to use a fixed color palette or generate random colors.
byte[] ColorPalette { get; set; } Custom color palette as RGB triplets [numColors, 3] with values in [0, 255].
double Alpha { get; set; } Alpha blending factor for overlay (0 = fully transparent, 1 = fully opaque).
double MinDisplayConfidence { get; set; } Minimum confidence threshold for displaying instances.
int ContourThickness { get; set; } Contour thickness in pixels.

Methods

Method Summary
byte[] GetADE20KPalette() Gets the default ADE20K color palette (150 classes).
byte[] GetCOCOPalette() Gets the default COCO panoptic color palette (133 classes).
byte[] GetCityscapesPalette() Gets a Cityscapes-style color palette (19 classes).