What it does

Root Image Preprocessor turns root scans into editable black-on-white masks in the browser. You can work through a small batch, trace a polygon region of interest, preview background cleanup, apply high-pass enhancement inside the region, correct the mask with a brush, and export a PNG.

Local workflow

Image decoding, processing, editing, and export happen in this browser tab. The app does not upload root scans to a server.

Before you start

• Use JPG or PNG scans with even illumination and clear contrast between roots and background.
• You can open up to 6 images at once. Each file must be no larger than 24 MB.
• Images are scaled so their longest edge is at most 1800 pixels, and the open batch is limited to 9 million decoded pixels.
• Save each result before closing or refreshing the tab; images and edit history exist only in memory.
• Treat the output as a preprocessing mask that still requires visual quality control.

Quick workflow

1. Select Upload images and choose one or more scans.
2. Select an image in Batch.
3. In Polygon Mode, click around the root region on the right canvas, then select Close Polygon.
4. Optionally tune Background threshold and Noise kernel, then select Preview Background Cleanup to inspect a separate cleanup preview.
5. Tune Blur radius and ROI threshold, then select Run ROI Processing.
6. In Manual Brush, choose Draw (black) to restore roots or Erase (white) to remove artifacts. Adjust Brush size as needed.
7. Use Undo Brush Stroke for recent corrections and select Download Processed PNG when the mask is ready.
8. Repeat for the remaining batch items, or use Clear batch to remove every open image from memory.

Controls & outputs

Control	Current behaviour
Upload images	Opens JPG/PNG scans within the file, batch, and decoded-pixel limits.
Batch	Switches between open images and shows Pending or Processed.
Background threshold	Sets the grayscale cutoff for the background-cleanup preview.
Noise kernel	Applies morphological opening to the background-cleanup preview.
Preview Background Cleanup	Updates the preview canvas; it does not replace the final ROI-processing input.
Blur radius	Sets the local blur used by final high-pass enhancement.
ROI threshold	Converts normalized high-pass response into the final binary mask.
Close / Undo Point / Reset Polygon	Completes, shortens, or clears the polygon. A polygon needs at least 3 points and supports at most 60.
Run ROI Processing	Creates a white-background, black-root mask inside the closed polygon.
Polygon Mode / Manual Brush	Switches between ROI definition and mask editing. Manual mode requires a processed result.
Draw / Erase / Brush size	Paints black or white strokes onto the processed mask.
Undo Brush Stroke	Restores a recent mask snapshot; the in-memory history keeps up to 4 snapshots.
Reset Processed Result	Clears processing and brush history but retains the current polygon until Reset Polygon is used.
Download Processed PNG	Exports the current processed mask as original-name-processed.png.

How it works

The browser decodes each image and reduces oversized dimensions before placing pixel data in memory. Preview Background Cleanup converts the image to grayscale, thresholds it, and optionally applies a morphological opening; it is a diagnostic preview controlled by Background threshold and Noise kernel.

Final Run ROI Processing follows a separate path. It blurs the original image, compares blurred and original grayscale intensity, normalizes that high-pass response, applies ROI threshold, and writes black detections only inside the polygon. The manual brush then edits that binary result directly.

This is a lightweight Canvas implementation inspired by a root-processing workflow, not an exact browser port of OpenCV scripts or a validated segmentation model.

Data, privacy & external services

• Files remain in the current browser tab and are not transmitted by the app.
• No account, cloud project, autosave, or recovery service is used.
• Clear batch, refresh, navigation, or tab closure discards the in-memory images and edit history.
• PNG downloads are created locally with the Canvas API.

Limitations

Research use

The generated PNG is a candidate mask, not a ground-truth root measurement. Inspect thin roots, crossings, scanner artifacts, and polygon edges before using it in ImageJ or another measurement workflow.

• Drag-and-drop is not implemented; use Upload images.
• Background-preview controls do not affect the final ROI high-pass mask.
• Large images and blur radii can block the browser while pixel operations run on the main thread.
• Automatic scaling reduces fine detail in high-resolution scans.
• Undo is intentionally shallow to limit memory use.
• There is no zoom tool, automatic root topology analysis, batch export, or direct WinRHIZO integration.

Troubleshooting

Problem	What to check
A file is rejected	Use JPG/PNG, keep it below 24 MB, and stay within the 6-file and 9-million-pixel batch limits.
Run ROI Processing does nothing	Select an image, add at least 3 polygon points, and select Close Polygon first.
Thin roots disappear	Reduce ROI threshold or Blur radius, rerun processing, then restore isolated details with Draw (black).
Noise remains	Increase ROI threshold, tighten the polygon, or remove artifacts with Erase (white).
Polygon editing is blocked	Select Reset Processed Result; use Reset Polygon as well if you need a new boundary.
Brush editing is unavailable	Run ROI processing before switching to Manual Brush.
The browser becomes slow	Reduce the source dimensions, open fewer images, or use a smaller blur radius.

Open Root Image Preprocessor

Back to App Lab