Products · Image segmentation

Generic Tree Ring Boundary Detector

Upload a wood cross-section image to visualize boundary pixels predicted by the generic model across varied species and capture conditions.

Back to all products

Live model trial

Upload an image

Choose one clear wood cross-section image. The generic model will resize it to 1504 × 1504 pixels and highlight likely tree-ring boundary pixels.

Example input

What to upload

Choose a focused image of the complete wood cross-section. Center the sample, keep the growth rings visible, and use even lighting with as little background clutter as possible.

Research prototype

This public trial exposes a neural boundary-segmentation model. It highlights likely ring boundaries but does not count annual rings or run pith-dependent geometric analysis.

Trial settings

Model: Generic tree-ring model
Model input: 1504 × 1504 px
Threshold: 0.50
Tiling: Disabled
Selected image: Up to 50 MB

Current limitations

Species, surface preparation, lighting, scale, background, image sharpness, and differences from the training data can affect the overlay. Use a centered, clearly visible cross-section and treat the result as an experimental aid.

Image handling

A resized and compressed copy is sent to the MSU Wood AI prediction service for this request. Your original stays on your device, and the lab application does not save the copy after the request completes.

Supported files

Use a clear JPEG, PNG, or WebP image up to 50 MB. Your browser prepares a copy no larger than 1504 pixels per side and under 4 MB before upload.