Contrast Equalization

Class: ContrastEqualizationBlockV1

Source: inference.core.workflows.core_steps.classical_cv.contrast_equalization.v1.ContrastEqualizationBlockV1

Enhance image contrast using configurable equalization methods (Contrast Stretching, Histogram Equalization, or Adaptive Equalization) to improve image visibility, distribute pixel intensities more evenly, and enhance details in low-contrast or poorly lit images for preprocessing, enhancement, and quality improvement workflows.

How This Block Works

This block enhances image contrast by redistributing pixel intensities using one of three equalization methods. The block:

1. Receives an input image to enhance with contrast equalization
2. Selects the contrast equalization method based on equalization_type parameter
3. Applies the selected equalization method:

For Contrast Stretching: - Calculates the 2nd and 98th percentiles of pixel intensities in the image (finds the darkest and brightest meaningful values, ignoring extreme outliers) - Stretches the intensity range between these percentiles to span the full 0-255 range - Enhances contrast by expanding the dynamic range while preserving relative intensity relationships - Useful for images with a narrow intensity range that need stretching to full range

For Histogram Equalization: - Normalizes pixel intensities to 0-1 range for processing - Computes and equalizes the image histogram to create a uniform distribution of pixel intensities - Redistributes pixel values so that each intensity level has approximately equal frequency - Scales the equalized values back to 0-255 range - Enhances contrast globally across the entire image, improving visibility of features

For Adaptive Equalization: - Normalizes pixel intensities to 0-1 range for processing - Applies adaptive histogram equalization (CLAHE - Contrast Limited Adaptive Histogram Equalization) - Divides the image into small regions and equalizes each region independently - Uses clip_limit=0.03 to limit contrast enhancement and prevent over-amplification of noise - Combines local equalized regions using bilinear interpolation for smooth transitions - Scales the result back to 0-255 range - Enhances contrast adaptively, preserving local details while improving overall visibility

1. Preserves image metadata from the original image
2. Returns the enhanced image with improved contrast

The block provides three methods with different characteristics: Contrast Stretching expands intensity ranges linearly, Histogram Equalization creates uniform intensity distribution globally, and Adaptive Equalization enhances contrast locally while preventing over-amplification. Each method works best for different scenarios - Contrast Stretching for images with narrow intensity ranges, Histogram Equalization for overall contrast improvement, and Adaptive Equalization for images with varying contrast across regions.

Common Use Cases

• Image Preprocessing for Models: Enhance image contrast before feeding to detection or classification models (e.g., improve contrast before object detection, enhance visibility before classification, prepare images for model processing), enabling improved model performance workflows
• Low-Contrast Image Enhancement: Improve visibility and details in low-contrast or poorly lit images (e.g., enhance dark images, improve visibility in low-light conditions, reveal details in low-contrast scenes), enabling image enhancement workflows
• Detail Enhancement: Reveal hidden details in images with poor contrast (e.g., enhance details in shadow regions, reveal features in dark areas, improve visibility of subtle details), enabling detail enhancement workflows
• Image Quality Improvement: Improve overall image quality and visibility (e.g., enhance overall image quality, improve visibility for analysis, optimize images for display), enabling image quality workflows
• Medical and Scientific Imaging: Enhance contrast in medical or scientific images for better analysis (e.g., enhance medical imaging contrast, improve scientific image visibility, prepare images for analysis), enabling scientific imaging workflows
• Document Image Enhancement: Improve contrast in scanned documents or document images (e.g., enhance document contrast, improve text visibility, optimize scanned documents), enabling document enhancement workflows

Connecting to Other Blocks

This block receives an image and produces an enhanced image with improved contrast:

• After image input blocks to enhance input images before further processing (e.g., enhance contrast in camera feeds, improve visibility in image inputs, optimize images for workflow processing), enabling image enhancement workflows
• Before detection or classification models to improve model performance with better contrast (e.g., enhance images before object detection, improve visibility for classification models, prepare images for model analysis), enabling enhanced model input workflows
• After preprocessing blocks to apply contrast enhancement after other preprocessing (e.g., enhance contrast after filtering, improve visibility after transformations, optimize images after preprocessing), enabling multi-stage enhancement workflows
• Before visualization blocks to display enhanced images with better visibility (e.g., visualize enhanced images, display improved contrast results, show enhancement effects), enabling enhanced visualization workflows
• Before analysis blocks that benefit from improved contrast (e.g., analyze enhanced images, process improved visibility images, work with optimized contrast), enabling enhanced analysis workflows
• In image quality improvement pipelines where contrast enhancement is part of a larger enhancement workflow (e.g., enhance images in multi-stage pipelines, improve quality through enhancement steps, optimize images in processing chains), enabling image quality pipeline workflows

Type identifier

Use the following identifier in step "type" field: roboflow_core/contrast_equalization@v1to add the block as as step in your workflow.

Properties

Name	Type	Description	Refs
name	str	Enter a unique identifier for this step..	❌
equalization_type	str	Type of contrast equalization method to apply: 'Contrast Stretching' stretches the intensity range between 2nd and 98th percentiles to full 0-255 range (linear expansion, good for narrow intensity ranges), 'Histogram Equalization' (default) creates uniform intensity distribution globally (equalizes histogram across entire image, good for overall contrast improvement), or 'Adaptive Equalization' enhances contrast locally in small regions while limiting over-amplification (CLAHE with clip_limit=0.03, good for images with varying contrast). Default is 'Histogram Equalization' which provides good general-purpose contrast enhancement. Choose based on image characteristics and enhancement needs..	✅

The Refs column marks possibility to parametrise the property with dynamic values available in workflow runtime. See Bindings for more info.

Available Connections

Compatible Blocks

Check what blocks you can connect to Contrast Equalization in version v1.

• inputs: Corner Visualization, Image Convert Grayscale, Label Visualization, Image Slicer, Image Blur, CSV Formatter, SIFT Comparison, Florence-2 Model, Google Gemini, OCR Model, Ellipse Visualization, Halo Visualization, Single-Label Classification Model, Webhook Sink, Contrast Equalization, Stability AI Outpainting, Camera Focus, Model Comparison Visualization, Stitch Images, Polygon Visualization, Object Detection Model, Stability AI Inpainting, Reference Path Visualization, OpenAI, OpenAI, Slack Notification, Circle Visualization, Background Subtraction, Stability AI Image Generation, Roboflow Dataset Upload, Icon Visualization, LMM For Classification, VLM as Classifier, Twilio SMS/MMS Notification, Model Monitoring Inference Aggregator, Color Visualization, Clip Comparison, Mask Visualization, Roboflow Dataset Upload, Anthropic Claude, Image Slicer, Pixelate Visualization, OpenAI, Email Notification, Image Contours, Google Gemini, Text Display, Blur Visualization, Stitch OCR Detections, Roboflow Custom Metadata, Triangle Visualization, Google Vision OCR, Relative Static Crop, Camera Focus, Classification Label Visualization, Multi-Label Classification Model, Image Threshold, LMM, Camera Calibration, Dot Visualization, Anthropic Claude, Background Color Visualization, Stitch OCR Detections, Polygon Zone Visualization, Keypoint Visualization, Grid Visualization, Dynamic Crop, Anthropic Claude, Keypoint Detection Model, Trace Visualization, Crop Visualization, Absolute Static Crop, Line Counter Visualization, Florence-2 Model, Google Gemini, Twilio SMS Notification, Image Preprocessing, Instance Segmentation Model, SIFT, Perspective Correction, Email Notification, Halo Visualization, EasyOCR, Local File Sink, Depth Estimation, CogVLM, Morphological Transformation, Polygon Visualization, OpenAI, QR Code Generator, Llama 3.2 Vision, VLM as Detector, Bounding Box Visualization
• outputs: Corner Visualization, Image Convert Grayscale, Label Visualization, Clip Comparison, Image Slicer, SmolVLM2, Image Blur, Florence-2 Model, SIFT Comparison, Google Gemini, OCR Model, Ellipse Visualization, Halo Visualization, Single-Label Classification Model, Stability AI Outpainting, Contrast Equalization, Perception Encoder Embedding Model, Qwen3-VL, Camera Focus, Stitch Images, Model Comparison Visualization, Polygon Visualization, Object Detection Model, Stability AI Inpainting, Reference Path Visualization, OpenAI, Detections Stabilizer, OpenAI, Circle Visualization, Background Subtraction, Roboflow Dataset Upload, Stability AI Image Generation, Icon Visualization, LMM For Classification, YOLO-World Model, VLM as Classifier, Pixel Color Count, Twilio SMS/MMS Notification, Object Detection Model, Multi-Label Classification Model, Color Visualization, Clip Comparison, SAM 3, Barcode Detection, Mask Visualization, Roboflow Dataset Upload, Anthropic Claude, Image Slicer, Template Matching, Buffer, Pixelate Visualization, OpenAI, Byte Tracker, CLIP Embedding Model, Instance Segmentation Model, Keypoint Detection Model, Email Notification, Image Contours, Google Gemini, Text Display, Blur Visualization, Triangle Visualization, Google Vision OCR, VLM as Detector, Relative Static Crop, Llama 3.2 Vision, Camera Focus, SAM 3, Classification Label Visualization, Multi-Label Classification Model, Image Threshold, LMM, Dot Visualization, Anthropic Claude, Camera Calibration, Background Color Visualization, Qwen2.5-VL, Dominant Color, Seg Preview, Polygon Zone Visualization, Keypoint Visualization, Anthropic Claude, Dynamic Crop, Keypoint Detection Model, Trace Visualization, SAM 3, Crop Visualization, Absolute Static Crop, Line Counter Visualization, Florence-2 Model, Google Gemini, Time in Zone, Detections Stitch, Segment Anything 2 Model, Moondream2, Image Preprocessing, Gaze Detection, Instance Segmentation Model, SIFT, Perspective Correction, Motion Detection, Halo Visualization, VLM as Classifier, EasyOCR, Depth Estimation, CogVLM, Morphological Transformation, OpenAI, Polygon Visualization, Single-Label Classification Model, QR Code Detection, VLM as Detector, Bounding Box Visualization

Input and Output Bindings

The available connections depend on its binding kinds. Check what binding kinds Contrast Equalization in version v1 has.

Bindings

• input

  • image (image): Input image to enhance with contrast equalization. The block applies one of three contrast equalization methods based on the equalization_type parameter. Works on color or grayscale images. The enhanced image will have improved contrast, better visibility, and enhanced details. Original image metadata is preserved in the output..
  • equalization_type (string): Type of contrast equalization method to apply: 'Contrast Stretching' stretches the intensity range between 2nd and 98th percentiles to full 0-255 range (linear expansion, good for narrow intensity ranges), 'Histogram Equalization' (default) creates uniform intensity distribution globally (equalizes histogram across entire image, good for overall contrast improvement), or 'Adaptive Equalization' enhances contrast locally in small regions while limiting over-amplification (CLAHE with clip_limit=0.03, good for images with varying contrast). Default is 'Histogram Equalization' which provides good general-purpose contrast enhancement. Choose based on image characteristics and enhancement needs..

• output

  • image (image): Image in workflows.

Example JSON definition of step Contrast Equalization in version v1

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/contrast_equalization@v1",
    "image": "$inputs.image",
    "equalization_type": "Histogram Equalization"
}