StrataQuest Glossary | ScientiaLux

Suppose a faint membrane is broken by grainy noise. An ordinary blur smears it into the background. Anisotropic diffusion smooths along the line and barely across it, so the grain fades and the gaps close.

Basic Operations Module

The arithmetic family does math one pixel at a time. Its main use is to reshape intensity for a more selective method downstream.

Basic Operations Module

Pixel math comes in three shapes: two images combined, one image with a constant, or one image passed through a curve such as a logarithm.

Basic Operations Module

Assign classes to objects records, for each detected event, the id of the class or proximity area it sits in. That ties Classifier and Proximity Areas output to per-cell data you can count.

Measurement & Quantification

With a square kernel, the box mean: convolution with uniform weights, the simplest and cheapest smoother (StrataQuest's default shape is Disk). Gauss weights neighbors by distance and keeps more structure for the same smoothing.

Basic Operations Module

Background Removal is a StrataQuest Pre-Processing engine. It estimates a background model for one grayscale channel and subtracts it, lowering a high background. The effect shows most in 16-bit images.

Image Processing & Correction

Averaging the brightness of stained pixels averages the wrong quantity. Transmitted light falls exponentially with stain concentration, so quantitative IHC and any absorption-based stain analysis converts to OD before computing averages.

Image Data Fundamentals

Take a pixel just inside a nucleus. A plain blur averages in the dark background beyond the edge. The Bilateral filter discounts neighbors of different brightness, so regions smooth and boundaries hold.

Basic Operations Module

Take two masks. One marks pixels inside a tissue ROI; the other marks pixels above the staining threshold. AND keeps the pixels in both. Four functions cover every way two masks combine into a third.

Basic Operations Module

Brightfield shines white light through stained tissue and records an RGB image. In StrataQuest it is the foundation for IHC analysis, where enzyme-linked chromogenic dyes such as DAB, hematoxylin and Fast Red make proteins visible.

Imaging Modalities & Platforms

Threshold a gradient image and boundaries come out thick and broken. The Canny edge detector thins each one to its crest and keeps weak stretches only where they connect to strong ones.

Basic Operations Module

An image is rarely one layer of pixel values. A brightfield color image has three; a fluorescence acquisition has one per reflector, often more. Each layer is a channel, a separate grayscale image.

Image Data Fundamentals

The Classifier is a Pre-Processing engine trained on areas you annotate. It sorts blocks of pixels into classes such as tumor, stroma and necrosis, producing a class map and a mask for each class.

Image Processing & Correction

StrataQuest stores detection results as coded images, in which every pixel of a detected object (event) carries that object's ID. Each holds up to 65,535 labels per Field of View; stitching links objects across fields.

Detection & Segmentation

Color Separation is a Pre-Processing engine for brightfield projects. It turns a color image into one grayscale image per marker, by unmixing against one reference shade per marker or by classifying pixels against several shades.

Image Processing & Correction

This group converts images between color spaces: BGR↔HLS, BGR↔HSV, BGR↔Lab, RGB↔Optical Density (OD), and RGB to Grayscale. The first three put color and brightness on separate axes, so each can be processed on its own.

Basic Operations Module

Color is three numbers per pixel, in any of several coordinate systems. The nine BOM color-conversion engines move between them, and a threshold or classification that works poorly in one space can work well in another.

Basic Operations Module

A black-and-white mask tells you which pixels are foreground, not which of them belong together. Labeling answers that question, turning one anonymous mask into separate objects, each with its own number.

Basic Operations Module

Convolution is the operation underlying most image filters that aren't rank-based. Slide a small grid across the image, multiply, sum. Blur, sharpening and edge detection differ only in the numbers in that grid.

Basic Operations Module

Every named convolutional filter in the filter family is a fixed grid of weights. The Custom filter lets you write the grid yourself, for when no named entry matches the structure you want to surface.

Basic Operations Module

Cutoffs are threshold lines on a diagram's axes. Each divides events into positive and negative along one measurement, making four quadrants in a scattergram and left and right regions in a histogram.

Data Visualization & Mining

Tissue Biology & Applications

Nuclei (Deep Learning) is a Detection engine that segments nuclei with a trained deep learning model. Like Nuclei, it needs fluorescence-like images and returns a coded image, one ID per nucleus.

Detection & Segmentation

Density of Events, a pre-processing engine, sums a Gaussian kernel over the positions of detected events. Its density map shows where tissue is crowded and where it is sparse, and quantifies local cellularity.

Measurement & Quantification

Derived Measurements makes new per-event measurements from existing ones with basic arithmetic. Each one combines two measurements, or a measurement and a value you enter, such as a pixel size.

Measurement & Quantification

StrataQuest is the quantitative analysis platform among TissueGnostics' digital pathology tools. It turns whole-slide images into single-cell data you can measure and reproduce.

Tissue Biology & Applications

A distance map stores in each pixel how far it lies from a structure. StrataQuest has three such engines: two measure outward from areas of interest, and one measures inward from event contours.

Spatial Analysis & Phenotyping

Dot Measurements counts and measures the dots inside each master event, such as a nucleus. It reports dot count, mean and summed dot intensity, total dot area, and per-dot lists of intensity and area.

Measurement & Quantification

A detection engine for small punctate signals in FISH (fluorescence in situ hybridization) and CISH (chromogenic in situ hybridization) images. It outputs each dot as an event, which Dot Measurements then counts inside each nucleus.

Detection & Segmentation

An engine is one step of a StrataQuest analysis, an algorithm that takes inputs and produces outputs. One engine adds two images; another runs classification. Layers arrange engines into complete analyses.

Detection & Segmentation

A field of view (FOV) is one camera frame of a scanned region. TissueFAXS acquires a region as a grid of overlapping FOVs, and StrataQuest runs most engines one FOV at a time.

Imaging Modalities & Platforms

Fill Holes resembles a Close but is a different operation. It tests whether each background pixel can reach the image edge without crossing foreground, and every pixel that cannot becomes foreground.

Basic Operations Module

Every filter in StrataQuest's Basic Operations Module, side by side: what each computes, its main settings, and which to reach for when you want to smooth, find edges, or lower a high background.

Basic Operations Module

Light of specific wavelengths excites labels on target molecules, and the emitted light is recorded channel by channel. StrataQuest analyzes multiple channels, each with an adjustable 16-bit dynamic range, pseudo-coloring and its own intensity scaling.

Imaging Modalities & Platforms

A binary mask labels every pixel as foreground or background. The foreground is the thing you want to count, measure or analyze. The background is everything else.

Image Data Fundamentals

Suppose one detection found the large nuclei and a second, with other settings, found the small ones. Fuse events gives you one coded image that holds both, with Image 1 on top.

Basic Operations Module

A gate is a shape drawn on a diagram, such as a histogram, scattergram, t-SNE, UMAP, SONG or phenotype diagram, that selects a set of events for further analysis.

Data Visualization & Mining

Blur an image slightly and pixel-level fluctuation settles while large-scale structures stay. In the BOM, the Gauss filter is the usual way to do this before any operation sensitive to pixel-level noise.

Basic Operations Module

A grayscale image gives each pixel a single intensity value and nothing else, with no color and no separate channels. It is the canonical input shape for most StrataQuest analysis engines.

Image Data Fundamentals

Grow (global) is a Measurements Masks engine that reshapes detected events into new measurement masks. It inflates them, deflates them or draws rings around their contours, so you can measure beyond or inside an event.

Detection & Segmentation

Tissue Biology & Applications

In StrataQuest, Histogram is also a diagram type that counts events, not pixels, across bins of one measurement. Its shape shows whether cells split into positive and negative, run as a gradient, or form sub-populations.

Data Visualization & Mining

Bit depth sets how many distinct grayscale values a pixel can hold. Sensors typically capture 12 or 16 bits; some downstream tools assume 8. Converting is sometimes harmless and sometimes loses real signal.

Image Data Fundamentals

Many BOM operations work on one channel of a multi-channel image at a time. Four operations in the Image Data group determine which channel goes where, and none of them changes an intensity value.

Basic Operations Module

Sometimes the next operation requires 8 bits and you have 16, or one channel and you have three. These six utilities reshape the data so that it fits.

Basic Operations Module

Most image-analysis pipelines start by filtering. A filter recomputes each pixel from its neighborhood, and filters differ in how they combine the values of that neighborhood.

Basic Operations Module

Each fluorophore, when excited, emits light at its own characteristic wavelength. That is what lets several markers share one section, and StrataQuest analyzes each one as an independent measurement channel.

Tissue Biology & Applications

IHC marks a protein with a colored precipitate visible in brightfield, and it is the most widely used protein stain on tissue sections. StrataQuest separates the stains with Color Separation and measures them per cell.

Tissue Biology & Applications

Labeling numbers every object, and the shapes it produces are a starting point. Two glands may share a thin bridge, and an object may carry a hole. These operations refine coded images after labeling.

Basic Operations Module

The Laplace filter responds strongly where the intensity surface curves sharply: at an edge, a thin line or a bright dot. Where intensity is locally smooth, it gives nearly zero. Its zero-crossings trace the edges.

Basic Operations Module

A layer is StrataQuest's unit of analysis, and its grammar: a set of engines, arranged in six categories, that detects one type of object and measures it.

Detection & Segmentation

Suppose you have a tissue mask and a necrosis mask. Tissue AND NOT necrosis is the viable tissue. Logical Operations build masks like that one, pixel by pixel, from AND, OR, NOT and XOR.

Basic Operations Module

A look-up table (LUT) maps each grayscale level to a display color. Fluorescence camera data is monochrome; LUTs like green (GFP), red (RFP), cyan (CFP), Fire and Ice add false color without changing the data.

Image Data Fundamentals

This post-detection engine lets you edit a segmentation by hand with four operations: Add for false negatives, Remove for false positives, Split for under-segmentation and Merge for over-segmentation.

Detection & Segmentation

Most engines treat a mask's foreground as the part to process, but the mask itself is neutral. Whether it surfaces a property or suppresses one is an interpretation, not a property of the data.

Image Data Fundamentals

Each pixel's neighborhood holds a vote, and the middle value wins. A lone extreme pixel is always outvoted, while a step edge passes through, because each side of the edge wins its own vote.

Basic Operations Module

Membrane is a detection engine that finds thin, line-shaped structures in a membrane-marker image and returns them twice: at their detected width, and as a one-pixel-wide centerline.

Detection & Segmentation

The Membrane detector finds membranes by how they look in an image: thin lines, darker or brighter than their surroundings, running in the directions you set.

Basic Operations Module

Membrane Measurements, in local and global versions, measures the membrane mask around each event: how far around it reaches and how far away it lies. The global version adds coverage, intensity and thickness.

Measurement & Quantification

Slide a small shape, the structuring element, over a mask. Erode keeps the points where it fits inside the foreground; Dilate adds every point it touches. Open, Close and Top-hat are built from those two.

Basic Operations Module

Multiplex IF labels several proteins at once on one tissue section, each with its own fluorophore. StrataQuest handles it with spectral unmixing, multi-channel measurement and hierarchical gating.

Tissue Biology & Applications

Nuclei, a Detection engine, finds and outlines each nucleus in a fluorescence-like image, where nuclei are bright on a dark background. Its result is a coded image with one ID per nucleus.

Detection & Segmentation

Stain absorbs light, so heavily stained tissue looks dark, but pixel values fall off exponentially as stain builds up. StrataQuest's two optical density operations convert between RGB and OD, a logarithmic scale that tracks stain.

Basic Operations Module

The Otsu Threshold (global) engine computes a single optimal intensity threshold for the whole sample. It uses Otsu's method, which separates foreground from background in a two-peaked histogram by minimizing the variance within each class.

Image Processing & Correction

Phenotype Interactions finds, for each cell of one phenotype, the cells of a second phenotype within a distance interval you set. It counts them, measures how far they are, and can draw each connection.

Spatial Analysis & Phenotyping

Phenotypes are named cell populations defined by marker combinations, such as CD3+CD8+ cytotoxic T cells. In StrataQuest they come out of gating in diagrams, and they feed reports, other diagrams and Phenotype Interactions.

Data Visualization & Mining

Suppose you want one image that is bright wherever either of two markers is. Max (every pixel) gives you that. Min (every pixel) gives the reverse: an image bright only where both are.

Basic Operations Module

The Post Processing group of the Basic Operations Module deletes events. Two operations test area, two test mean intensity, and two test overlap with a seed image. Mask cleanup lives in the Morphological group.

Basic Operations Module

Projection is a Pre-Processing engine that fuses selected images, usually the slices of a z-stack, into one image with greater depth of field. It offers nine methods: Sum, Max, Min, Average and five wavelets.

Image Processing & Correction

Proximity Areas, a StrataQuest pre-processing engine, draws concentric bands at chosen distances around reference structures, such as 0–50, 50–100 and 100–200 µm from a tumor border. Then measure any cell population band by band.

Spatial Analysis & Phenotyping

Raw Data shows the measurements of a layer's detected events as a table: one row per event, one column per measurement.

Measurement & Quantification

After removal or other label edits, a coded image can be numbered 1, 2, 5, 7, 12, or carry one ID on separate regions. The Engines Reference recommends running Re-Label to correct the labels list.

Basic Operations Module

An ROI is an area you define on the sample: a rectangle, an ellipse, a custom polygon or a freedrawn shape. It yields region-specific statistics, and StrataQuest skips computation outside it.

Spatial Analysis & Phenotyping

Detection writes every event it finds to a coded image. The analysis usually requires a subset: events large enough, bright enough, or touched by a seed. Six BOM operations select it.

Basic Operations Module

Two Post-Processing engines prune a layer's detected events. Remove objects v2 keeps events inside measurement ranges such as Area and Mean Intensity; Remove objects touching border drops events at the sample or region border.

Detection & Segmentation

The Scaling factor appears on the Basic Operations Module's arithmetic operations: Add images, Add weighted images, Multiply with value, Polynomial and the rest. Each multiplies its result by 2 raised to the Scaling factor.

Image Data Fundamentals

The scattergram is StrataQuest's main data view: a 2D plot with every detected event placed by two measurements. It shows population structure and co-expression, and it is where you draw gates and place cutoffs.

Data Visualization & Mining

The Sharp filter raises the contrast of fine detail and can improve a slightly out-of-focus image. Its one setting is Kernel radius (px). The Unsharp filter, despite its name, does something else.

Basic Operations Module

Thin a binary mask of membranes or fibers to one-pixel centerlines, and shape becomes a network of end points, branch points and the lines between. Seven more operations then find, prune, extend or split it.

Basic Operations Module

Walk across a nucleus in a DAPI image and the brightness climbs, then falls. StrataQuest's three Sobel filters mark where it changes sharply: along the horizontal axis, the vertical axis, or the diagonal.

Basic Operations Module

Spatial phenotyping describes cells by their surroundings as well as their markers. StrataQuest has no engine by that name; you assemble it from Phenotype Interactions, Proximity Areas, Distance Maps and Density of Events.

Spatial Analysis & Phenotyping

Spectral Unmixing is a Pre-Processing engine for FL multispectral experiments. It compares each pixel's spectrum with reference spectra you define and writes one grayscale image per marker, solved by non-negative least squares.

Image Processing & Correction

Standard Measurements computes the measurements you select for each detected event: shape from a coded image, intensity from one grayscale image. It comes in a global version, for stitched events, and a local one.

Measurement & Quantification

Lay two images on top of each other and, at every pixel, keep only the brighter value, or only the dimmer one. That is the whole Statistics group.

Basic Operations Module

Stitching combines FOVs into one image. In TissueFAXS it finds the pixel shift between overlapping neighbors to correct stage errors; in StrataQuest, object stitching joins events that cross FOV borders.

Imaging Modalities & Platforms

The technical name is uglier than the concept. A structuring element is a small shape that morphology operations slide across a mask, and it plays the part a convolution kernel plays in filtering.

Basic Operations Module

An analysis is organized as layers of engines in six categories, built in the Layers Editor. Inside one engine, the Basic Operations Module, the Operations Editor combines simple operations.

Detection & Segmentation

Twelve operations test each pixel against fixed or automatic values. Most return a mask, 255 where the test passes; the Threshold operations instead replace passing values and return a grayscale image.

Basic Operations Module

Twelve Basic Operations Module operations compare each pixel with a value, an interval or an Otsu threshold. Compare and Otsu return a binary mask; Threshold caps values and returns an image.

Basic Operations Module

Tissue cytometry brings flow cytometry's tools, such as gating, scatter plots and histograms, to intact tissue sections. Each cell is measured where it lies, so its population and its position can be analyzed together.

Tissue Biology & Applications

Tissue Detection marks which parts of a slide's preview image hold tissue. As a Pre-Processing engine it returns a tissue mask; as the tool behind Autodetect ROIs it draws ROIs for you.

Detection & Segmentation

Top-hat subtracts a morphologically-opened copy of an image from the original. What remains is what the opening removed: bright structures smaller than the structuring element, free of the slowly varying background.

Basic Operations Module

Total Area, a Detection engine, thresholds a grayscale image and gives each connected area above the threshold its own ID. Measured with Standard Measurements, those area events tell you how much is positive.

Detection & Segmentation

A heavily blurred copy of an image keeps the slowly varying background and loses the fine detail. Subtract it from the original and the background drops, while nuclei and other small structures remain.

Basic Operations Module

Virtual Channel is a Pre-Processing engine that combines the grayscale images you select, original channels or outputs of other engines, into one image. It has one parameter: Select gray images.

Image Processing & Correction

Virtual slides are the main input to StrataQuest projects: whole-slide images stored as resolution pyramids. StrataQuest references them in place without copying them, and supports formats from TissueFAXS, Hamamatsu NanoZoomer, Mirax, BigTIFF and OME-TIFF.

Imaging Modalities & Platforms

Watershed (global), a Detection engine, floods a topographic map from the non-zero pixels of a Seeds image and builds barriers where floods meet. The barriers split connected regions into separate events.

Detection & Segmentation