How-to 04 · Scoring, offline

Fiji multi-point and Cell Counter

Where there is no CoralNet account, no connection, or you simply want the scoring in the same program as the frame check. Fiji's Cell Counter plugin was written for counting cells on a slide; a benthic class is just a counter type, and a frame is just a slide.

Get the points onto the frame

The points must be random and recorded, exactly as for the manual overlay. Two ways to put them in Fiji:

From a point list. Generate the CSV with the overlay script, then in Fiji open the frame and use File → Import → XY Coordinates… (reads a two-column text file of x and y) — or open the CSV as a Results table and Edit → Selection → Specify… is the long way. The multi-point tool selection that results is your 50 points. Save it: Edit → Selection → Add to Manager, then in the ROI Manager More → Save…. That .roi or .zip file is the point-position record.
Directly, with a random overlay image. Open the annotated copy the overlay script produced — the points are already drawn on it — and score that. Simpler, and the CSV it came with is the position record.

Do not place points by hand with the multi-point tool and call them random. The tool is for marking positions you already have; it is not a random-number source.

Counting with Cell Counter

Open the plugin: Plugins → Analyze → Cell Counter (in some Fiji builds it is under Plugins → Cell Counter). A window opens with numbered counter types and an Initialize button. Click Initialize with the frame active.
Rename the counter types to the course's classes — hard coral, soft coral, dead coral, macroalgae, coralline algae, turf algae, other invertebrates, sand, rubble, rock, indeterminate. Cell Counter allows renaming through its options; if your build does not, keep a key on paper: type 1 = hard coral, and so on. Write the key in the Entry.
Select a class, then click on each point that belongs to it. Work through the 50 points in order; at each, decide what is directly underneath, pick that class's counter, and click the point. The window keeps a running total per type. A point on bleached tissue is hard coral; a point the frame cannot support is indeterminate, never a guess.
Check the totals sum to 50 (or 25). If not, a point was skipped or double-counted — Delete removes the last mark of the selected type.
Export the results: Results in the Cell Counter window writes a table of counts per type; Export Markers saves the marked positions as an XML file. Save both beside the frame, named to match it.

The numbers

Counts over the point count are the cover figures — 16 of 50 on hard coral is 32 %. Read the sampling Standard Error from the table for that cover and point count and write both in Section 4. Fewer than three points in a class: report the count, not a percentage with a ±.

Your buddy scores the same points in their own Cell Counter session — Initialize again on a fresh copy of the frame, or after Reset — without seeing your marks. Two results files, two columns in Section 4, never averaged.

What goes in the archive

The camera's file, untouched.
The point-position file: the overlay CSV, the ROI set, or the exported markers XML — at least one that gives x and y for every point.
Both scorers' Cell Counter results.
In Section 12: "Fiji (ImageJ) <version>, Cell Counter, manual classification" as the scoring tool. No model, no model version.

One thing the plugin cannot do for you. It counts what you click. The rule that makes the count honest — indeterminate for a point the frame cannot support, hard coral for bleached tissue, no averaging with your buddy — is yours to apply at every click, and the lesson is where it is explained.

Last checked against Fiji 2.x (ImageJ 1.54) in September 2026. The Cell Counter plugin ships with Fiji; if it is missing from your build, Help → Update… restores it. Menu locations are as of that check; if a step no longer matches what you see, say so.