StrataQuest Home StrataQuest Glossary Region of Interest (ROI)
Image Processing & Correction

Region of Interest (ROI)

Scoping a question to a region — the analysis becomes about that region, not the whole image

Definition
A Region of Interest is a sub-region of an image — typically a rectangle, ellipse, polygon, or arbitrary maskLoading... — that scopes which pixels an analysis engine acts on. Pixels inside the ROI are processed normally; pixels outside are excluded from the engine's computation, leaving them unchanged in the output. ROIs are drawn manually (operator outlines a tumor region), computed automatically (a ClassifierLoading... produces a tissue-class mask used as ROI), or imported from co-registered annotations. Once defined, an ROI propagates through subsequent engines that accept an ROI parameter — a single ROI definition scopes the entire downstream chain.
User Interface - Image Viewer + Toolbar
Video · Primary
User Interface - How to use ROI, ANN and EA tool
Video · Primary
Supporting
H&EDSPTCH
Video · Supporting
Detection - Automatic Background Threshold params
Video · Supporting
ROI restricts the scope, not the engine
The same engine, applied only here
Manual, computed, or imported
Three sources, same downstream behavior
ROIs are masks under the hood
Same data structure, different role
ROI restriction often replaces brute-force segmentation
Why bother detecting tumor cells outside the tumor?
Share This Term
Term Connections