Trout Identification

Trout Identification

Reading the spot patterns on trout to identify individual fish — a non-invasive way to monitor populations.
Manual
  1. Image Selection

    Choose an image from the examples provided below, or upload your own data.

  2. Run the ML model

    Click the 'Identify' button to initiate the machine learning model.

  3. Visualize the results

    The system will standardize the fish images and compare them to our database of trout from British Columbia. It then identifies the individual fish that has the highest probability of matching. If no match is found, the fish is classified as a new entry.

Overview

Trout Identification reads the spot patterns on a trout the way a fingerprint scanner reads a thumb. Each fish carries a unique, stable arrangement of black spots that stays consistent throughout its life — so a clear photo is enough to recognize the same individual again. Developed with Lumax AI, the system was built around Westslope Cutthroat Trout monitored in the Elk River, British Columbia.

It compares a new image against a catalog of known fish and returns the most likely match — or flags the trout as a new entry when it hasn’t been seen before. No tagging, no handling, no stress on the animal.

How It Works

From a photo to a confident match:

Image input

Works from ordinary trout photographs — choose an example or upload your own image of the fish.

Standardization

Aligns and standardizes the fish so spot patterns can be compared consistently, regardless of how the photo was taken.

Spot-pattern matching

Matches the unique spot arrangement against a catalog of known trout and ranks the most likely individual.

New-entry detection

When no confident match exists, the fish is logged as a new individual — growing the catalog over time.

Why It Matters

An indicator species

Westslope Cutthroat Trout need cold, clean water, so their numbers are a sensitive gauge of freshwater ecosystem health.

Non-invasive

Identification from natural markings means no tags or handling — monitoring that doesn't disturb the fish.

Population tracking

Re-identifying individuals over time supports survival estimates, movement studies, and long-term population monitoring.

In partnership with

Learn more about the project

See the full trout identification project and our work with Lumax AI on the Elk River — the research behind this tool.

View the project