Measurement & Analysis¶

Blob Detect¶

Detects bright blobs using Laplacian-of-Gaussian (LoG) filtering.

Details

Finds roughly circular bright spots such as cells, nuclei, vesicles, or puncta using skimage.feature.blob_log.

Outputs:

table — one row per blob with columns y, x, radius_px
overlay — original image with detected blobs circled in red
Min Radius — smallest blob radius to detect (pixels).
Max Radius — largest blob radius to detect (pixels).
Threshold — detection sensitivity; lower values find more blobs.

Direction	Port	Type
Input	`image`	image
Output	`overlay`	image

Properties: Min Radius (px), Max Radius (px), Threshold

Colocalization¶

Computes colocalization metrics between two channels.

Details

All metrics respect the mask input when connected. Without a mask, all pixels are used.

Metrics:

Pearson r — linear correlation (-1 to 1)
Spearman r — rank correlation, robust to non-linear relationships
Kendall tau — rank correlation, more robust for small samples
MOC — Manders' Overlap Coefficient (0 to 1)
M1 — fraction of ch1 intensity where ch2 is above its Otsu threshold
M2 — fraction of ch2 intensity where ch1 is above its Otsu threshold
ICQ — Li's Intensity Correlation Quotient (-0.5 to 0.5)

Accepts an optional label image — when connected, metrics are computed per-label (one row per region) instead of for the whole image.

Plot options: Scatter, Regression (with fit line per label), or Bar Chart (Pearson r per label).

Direction	Port	Type
Input	`ch1`	image
Input	`ch2`	image
Input	`label_image`	label
Output	`table`	table
Output	`figure`	figure

Properties: Plot Type

GLCM Texture¶

Computes Haralick texture features from a Grey-Level Co-occurrence Matrix (GLCM).

Details

Averages texture features over four orientations (0, 45, 90, 135 degrees) at the given pixel distance. Outputs a single-row table with columns: contrast, dissimilarity, homogeneity, energy, correlation, ASM.

Distance — pixel offset for co-occurrence pairs.
Grey Levels — number of quantisation levels (fewer = faster, coarser).

Direction	Port	Type
Input	`image`	image
Output	`table`	table

Properties: Distance (px), Grey Levels

Image Stats¶

Measures whole-mask region properties and pixel intensity statistics in a single table row.

Details

Connect an image, a mask, or both -- at least one must be connected.

Mask columns (present when mask is connected):

image_size_px — total pixels in the image (H x W); denominator for area_fraction
area_px — number of foreground pixels
area_fraction — area_px / image_size_px (0-1); multiply by 100 for %
perimeter_px — boundary length in pixels
solidity — area / convex_hull_area (1 = convex)
eccentricity — shape elongation (0 = circle, 1 = line)
major_axis_px — major axis of the fitted ellipse
minor_axis_px — minor axis of the fitted ellipse
extent — area / bounding_box_area
euler_number — 1 = no holes; decreases by 1 per enclosed hole
centroid_y, centroid_x — pixel coordinates of the mask centroid

Intensity columns (present when image is connected, pixel values 0-255):

mean, std, min, max, median — overall grayscale or luminance; restricted to masked region when mask is also connected

Per-channel columns (RGB image with Per Channel checked):

mean_r/g/b, std_r/g/b, min_r/g/b, max_r/g/b
Column Prefix — optional string prepended to all column names.

Direction	Port	Type
Input	`image`	image
Input	`mask`	mask
Output	`table`	table

Properties: ``

Intensity Profile¶

Plots pixel intensity along an interactively drawn line segment.

Details

Draw the line directly on the image preview. The plot shows intensity (or per-channel R/G/B) vs distance in pixels. Useful for measuring gradients, checking membrane sharpness, or verifying stain distribution across tissue layers.

Direction	Port	Type
Input	`image`	image
Output	`figure`	figure

Properties: ``

Find Contours¶

Finds all contours in a binary mask or edge image at a given intensity level.

Details

Outputs:

mask — binary image with all selected contours drawn as lines
table — coordinate table with columns contour_id, x, y

Modes:

All contours — return every contour found, sorted largest first
Largest only — return only the contour with the greatest enclosed area
Filter by min area — discard contours with enclosed area below Min Area
Contour Level — intensity threshold for contour detection (normalised 0-1).
Line Width — stroke width in pixels for the output mask drawing.

Direction	Port	Type
Input	`image/mask`	image
Output	`mask`	mask
Output	`table`	table

Properties: Contour Level (0–1), Line Width (px), Mode, Min Area (px²)

Hough Circles¶

Detects circles in a Canny edge image using the Hough circle transform.

Details

Sweeps a range of radii and votes for circle centres; peaks in the accumulator become detections. Connect a CannyEdgeNode output to the input.

Outputs:

overlay — RGB image with detected circles drawn in green
table — columns cx, cy, radius for every detected circle
Min Radius — smallest circle radius to search for (pixels).
Max Radius — largest circle radius to search for (pixels).
Threshold — fraction of the peak accumulator value required for a detection.

Direction	Port	Type
Input	`mask`	mask
Output	`overlay`	image

Properties: Min Radius (px), Max Radius (px), Max Circles, Threshold (0–1)

Hough Lines¶

Detects straight lines in a Canny edge image using the Hough line transform.

Details

Each detected line is described by (theta, rho): the perpendicular angle and distance from the image origin. Lines are extended to the full image boundary for the overlay. Connect a CannyEdgeNode output to the input.

Outputs:

overlay — RGB image with detected lines drawn in red
table — columns theta (rad), rho (px), and endpoint coordinates x0, y0, x1, y1
Threshold — fraction of the peak accumulator value required for a detection.
Min Distance — minimum pixel separation between detected lines.
Min Angle — minimum angular separation in degrees between detected lines.

Direction	Port	Type
Input	`mask`	mask
Output	`overlay`	image

Properties: Max Lines, Threshold (0–1), Min Distance (px), Min Angle (deg)

Hough Ellipse¶

Detects ellipses in a Canny edge image using the Hough ellipse transform.

Details

Slow on large images -- resize input to under 300x300 px for best speed. Uses skimage.transform.hough_ellipse.

Outputs:

overlay — RGB image with detected ellipses drawn in cyan
table — columns cx, cy, a (semi-major), b (semi-minor), orientation (rad)
Min Semi-Major — smallest semi-major axis to search for (pixels).
Max Semi-Major — largest semi-major axis to search for (pixels).
Accuracy — step size in pixels for the accumulator; larger = faster but coarser.
Threshold — fraction of the peak accumulator value required for a detection.

Direction	Port	Type
Input	`mask`	mask
Output	`overlay`	image

Properties: Min Semi-Major (px), Max Semi-Major (px), Accuracy (px step), Max Ellipses, Threshold (0–1)

Image Histogram¶

Plots the pixel intensity histogram of an image.

Details

Outputs a figure showing intensity distribution per channel (R/G/B for colour images, a single Intensity curve for grayscale). Optionally accepts a mask to restrict the histogram to the masked region only. Also outputs a table with columns Pixel_Value and one column per channel.

Bins — number of histogram bins (default 256; auto-capped at max pixel value + 1).
Log Y-axis — show frequency on a log scale.
Fill Alpha — line / fill opacity (0.0-1.0).

Direction	Port	Type
Input	`image`	image
Input	`mask`	mask
Output	`figure`	figure
Output	`table`	table

Properties: Bins, `,Fill Alpha`

Save Image¶

Saves an image to disk. Click Browse to choose file location and format.

Details

Inputs:

image — ImageData to save

Supported formats: PNG, TIFF (16-bit preserved), JPEG. Users can also type any path with a custom extension directly. TIFF preserves bit depth and scale metadata when available.

Direction	Port	Type
Input	`image`	image