Image Processing

Image Processing

pupil covers 2D image processing: the building-block operations you compose before and after perception - bitwise math, cropping, geometric transforms, color conversion, enhancement, filtering, morphology, and measurement. Pick the group that matches the operation you need.

Bitwise Operations

Bitwise operations combine or invert images and binary masks pixel by pixel. Use them to intersect, union, subtract, or complement masks when composing detection and segmentation results.

Bitwise AND Images 0

Bitwise AND Images 1

Bitwise AND Images

Perform a pixel-wise bitwise AND operation between two images or binary masks, useful for mask intersection and logical filtering.

Bitwise Difference Images 0

Bitwise Difference Images 1

Bitwise Difference Images

Compute the pixel-wise absolute difference between two images to highlight changes, useful for motion detection, defect inspection, and image comparison.

Bitwise NOT Image 0

Bitwise NOT Image 1

Bitwise NOT Image

Invert an image or binary mask by performing a pixel-wise bitwise NOT operation. Useful for mask inversion, foreground/background swapping, and logical complement operations.

Bitwise OR Images 0

Bitwise OR Images 1

Bitwise OR Images

Combine two images or binary masks using pixel-wise bitwise OR operation. Useful for mask union, merging detection results, and filling gaps.

Bitwise XOR Images 0

Bitwise XOR Images 1

Bitwise XOR Images

Compute pixel-wise exclusive regions between two binary images using bitwise XOR. Useful for change detection, mask comparison, and highlighting differences.

Cropping & Padding

Crop an image down to a region of interest - centered, by bounding boxes, or by an arbitrary polygon - or pad it out to a target size. Use these to prepare fixed-size model inputs or to isolate detected objects.

Crop Image Center 0

Crop Image Center 1

Crop Image Center

Crop an image to specified dimensions centered on the image, with optional padding for smaller images. Useful for fixed-size model inputs, thumbnails, and normalization.

Crop Image Using Bounding Boxes 0

Crop Image Using Bounding Boxes 1

Crop Image Using Bounding Boxes

Crop an image using multiple rectangular bounding boxes, optionally retaining coordinate metadata. Useful for object detection, ROI extraction, and batch processing.

Crop Image Using Polygon 0

Crop Image Using Polygon 1

Crop Image Using Polygon

Crop an image using an arbitrary polygon mask, retaining pixels inside the polygon. Useful for irregular shapes, segmentation contours, and non-rectangular ROIs.

Pad Image

Add padding to an image on top, bottom, left, and right with configurable border handling and fill value.

Geometric Transforms

Resize, rotate, shift, or rescale an image while controlling aspect ratio and border handling. Pyramid up/down-sampling gives you clean multi-scale versions of the same image.

Resize Image

Resize an image by scale factor or explicit width and height for flexible scaling in vision pipelines.

Resize Image With Aspect Fit 0

Resize Image With Aspect Fit 1

Resize Image With Aspect Fit

Resize an image to fit within target dimensions while preserving aspect ratio, with optional padding.

Rotate Image

Rotate an image by a specified angle with options to keep original dimensions or expand canvas.

Translate Image 0

Translate Image 1

Translate Image

Shift an image by a fixed number of pixels in horizontal and vertical directions with configurable border handling and interpolation.

Transform Using Pyramid Downsampling 0

Transform Using Pyramid Downsampling 1

Transform Using Pyramid Downsampling

Downsample an image using Gaussian pyramid smoothing and subsampling for multi-scale processing and efficient image analysis.

Transform Using Pyramid Upsampling 0

Transform Using Pyramid Upsampling 1

Transform Using Pyramid Upsampling

Upsample an image using Gaussian pyramid smoothing and interpolation for multi-scale reconstruction and image enlargement.

Color & Channels

Convert between color spaces, split an image into its individual channels or merge channels back together, and blend two images with a weighted overlay.

Convert Image Color Space 0

Convert Image Color Space 1

Convert Image Color Space

Convert an image between color spaces (e.g., RGB, BGR, HSV, GRAY, RGBA) for segmentation, display, or cross-library compatibility.

Split Image Into Channels 0

Split Image Into Channels 1

Split Image Into Channels

Split a multi-channel image into individual color channels for per-channel processing and analysis.

Merge Image From Channels 0

Merge Image From Channels 1

Merge Image From Channels

Merge multiple single-channel images into a multi-channel image, useful for per-channel processing and color reconstruction.

Overlay Images Using Weighted Overlay 0

Overlay Images Using Weighted Overlay 1

Overlay Images Using Weighted Overlay

Blend two images using weighted overlay for visualization, crossfade, or multi-exposure effects.

Enhancement & Normalization

Correct exposure, contrast, and color casts, and normalize intensity so downstream steps see consistent images across varying lighting conditions.

Enhance Image Using Auto Gamma Correction 0

Enhance Image Using Auto Gamma Correction 1

Enhance Image Using Auto Gamma Correction

Automatically adjust image brightness using adaptive gamma estimation to normalize exposure. Ideal for varying lighting conditions.

Enhance Image Using CLAHE 0

Enhance Image Using CLAHE 1

Enhance Image Using CLAHE

Apply Contrast Limited Adaptive Histogram Equalization (CLAHE) to enhance local contrast while controlling noise amplification, ideal for low-light or unevenly illuminated images.

Enhance Image Using White Balance 0

Enhance Image Using White Balance 1

Enhance Image Using White Balance

Apply automatic white balance correction to remove color casts and adjust color temperature, ensuring neutral colors appear natural under varying illumination.

Normalize Image Intensity 0

Normalize Image Intensity 1

Normalize Image Intensity

Normalize image intensity values using minmax or histogram-based methods for consistent contrast and downstream processing.

Smoothing & Noise Reduction

Reduce noise before thresholding or edge detection. Choose fast averaging (box, blur), edge-preserving smoothing (bilateral, Gaussian), or impulse-noise removal (median).

Filter Image Using Bilateral 0

Filter Image Using Bilateral 1

Filter Image Using Bilateral

Apply a bilateral filter to reduce image noise while preserving edges by considering both spatial proximity and color similarity, ideal for edge-preserving smoothing in photography and vision pipelines.

Filter Image Using Blur 0

Filter Image Using Blur 1

Filter Image Using Blur

Apply a simple box blur to quickly smooth an image by averaging pixel values within a kernel. Ideal for fast preprocessing, background estimation, or artistic blur effects where edge preservation is not critical.

Filter Image Using Box 0

Filter Image Using Box 1

Filter Image Using Box

Apply a normalized box filter to perform basic averaging with control over kernel size, normalization, output depth, and border handling. Ideal for preprocessing, local averaging, and feature extraction where precision matters.

Filter Image Using Gaussian Blur 0

Filter Image Using Gaussian Blur 1

Filter Image Using Gaussian Blur

Apply a Gaussian blur to smoothly reduce noise while preserving edges, ideal for preprocessing, image enhancement, and feature extraction in computer vision pipelines.

Filter Image Using Median Blur 0

Filter Image Using Median Blur 1

Filter Image Using Median Blur

Apply a median blur filter to remove salt-and-pepper noise while preserving edges, ideal for impulse noise removal and image restoration in vision pipelines.

Edge Detection

Extract edges and gradients using first- and second-order derivative operators. Sobel and Scharr give directional gradients; the Laplacian catches fine detail and zero-crossings.

Filter Image Using Sobel 0

Filter Image Using Sobel 1

Filter Image Using Sobel

Apply Sobel filter for directional edge detection, computing gradients in X and Y directions to detect edges and orientation.

Filter Image Using Scharr 0

Filter Image Using Scharr 1

Filter Image Using Scharr

Apply Scharr filter for high-accuracy edge detection with improved rotation invariance, suitable for gradient-based applications.

Filter Image Using Laplacian Filter 0

Filter Image Using Laplacian Filter 1

Filter Image Using Laplacian Filter

Apply a Laplacian filter to detect edges and fine details using second-order derivatives, ideal for precise edge localization, image sharpening, and zero-crossing detection.

Ridge & Texture Filters

Enhance thin, tubular, or oriented structures - vessels, fibers, cracks, PCB traces, textures - using multi-scale ridge and texture filters.

Filter Image Using Frangi 0

Filter Image Using Frangi 1

Filter Image Using Frangi

Apply the Frangi vesselness filter to enhance tubular structures in images, ideal for medical imaging, retinal scans, angiography, and other applications requiring vessel detection.

Filter Image Using Hessian Filter 0

Filter Image Using Hessian Filter 1

Filter Image Using Hessian Filter

Apply a Hessian-based vesselness filter to detect tubular and ridge-like structures, offering a faster alternative to Frangi for vessel detection and ridge enhancement in biomedical and industrial imaging.

Filter Image Using Meijering 0

Filter Image Using Meijering 1

Filter Image Using Meijering

Apply the Meijering filter to detect neurites and fine branching structures in biomedical images, ideal for neuronal imaging and morphological analysis.

Filter Image Using Sato Filter 0

Filter Image Using Sato Filter 1

Filter Image Using Sato Filter

Apply Sato filter for multi-scale ridge detection to enhance thin linear structures such as fibers, vessels, cracks, and PCB traces.

Filter Image Using Gabor 0

Filter Image Using Gabor 1

Filter Image Using Gabor

Apply a Gabor filter to detect oriented textures and features at specific scales and orientations, ideal for texture analysis, fingerprint recognition, and biomedical imaging.

Morphological Operations

Reshape binary regions: grow or shrink them (dilate/erode), clean up noise and holes (open/close), extract boundaries or fine features (gradient, top-hat, black-hat), or reduce shapes to their skeleton (thinning).

Filter Image Using Morphological Dilation 0

Filter Image Using Morphological Dilation 1

Filter Image Using Morphological Dilation

Apply morphological dilation to expand bright regions and fill small gaps or holes, useful for mask expansion, object connection, and feature enhancement.

Filter Image Using Morphological Erosion 0

Filter Image Using Morphological Erosion 1

Filter Image Using Morphological Erosion

Apply morphological erosion to shrink bright regions and remove small noise, useful for binary image cleanup, object separation, and feature removal.

Filter Image Using Morphological Open 0

Filter Image Using Morphological Open 1

Filter Image Using Morphological Open

Apply morphological opening (erosion followed by dilation) to remove small noise while preserving object size, effective for binary image cleanup and segmentation refinement.

Filter Image Using Morphological Close 0

Filter Image Using Morphological Close 1

Filter Image Using Morphological Close

Apply morphological closing (dilation followed by erosion) to fill holes and connect nearby objects while preserving overall shape, ideal for segmentation cleanup and feature completion.

Filter Image Using Morphological Gradient 0

Filter Image Using Morphological Gradient 1

Filter Image Using Morphological Gradient

Apply morphological gradient to extract object boundaries by computing the difference between dilation and erosion, useful for robust edge detection in binary or grayscale images.

Filter Image Using Morphological Tophat Filter 0

Filter Image Using Morphological Tophat Filter 1

Filter Image Using Morphological Tophat Filter

Apply morphological top-hat transform to extract or enhance small bright features in images, useful for background correction, small object detection, and detail enhancement.

Filter Image Using Morphological Blackhat 0

Filter Image Using Morphological Blackhat 1

Filter Image Using Morphological Blackhat

Apply morphological black-hat transform to extract small dark features, cracks, or holes on bright backgrounds, ideal for defect and surface inspection.

Filter Image Using Morphological Thinning Filter 0

Filter Image Using Morphological Thinning Filter 1

Filter Image Using Morphological Thinning Filter

Apply skeletonization (thinning) to binary images to reduce objects to their skeletal structure while preserving connectivity and topology.

Measurement & Geometry

Measure properties of a mask - its centroid and principal axes - or project a pixel into 3D camera coordinates using camera intrinsics.

Calculate Image Centroid 0

Calculate Image Centroid 1

Calculate Image Centroid

Compute the centroid (center of mass) of non-zero pixels in a binary mask. Useful for object localization, alignment, and tracking.

Calculate Image PCA 0

Calculate Image PCA 1

Calculate Image PCA

Perform principal component analysis on a binary mask to compute centroid, eigenvectors, eigenvalues, and principal angle with visualization. Useful for shape orientation analysis, alignment, and morphological characterization.

Project Pixel to Camera Point 0

Project Pixel to Camera Point

Convert a 2D pixel coordinate and depth value into a 3D point in camera coordinates using camera intrinsics and distortion coefficients.

Where to Go Next?

Continue to the next tutorial.

Point Cloud Processing

3D point cloud filtering, downsampling, and processing with Vitreous.

Next tutorial →

Support

Join Discord

Ask questions, share what you're building, and get help from the community.

Report a bug?

Found a bug or something not working as expected? Raise an issue on our GitHub.

Raise an issue →