MUSHROOM VISION / TESTING ROUND 4
Look at the evidence.
Then help us improve it.
A saved comparison of original images, reduced-detail views and tiled scans. Every result below stays connected to the image and AI observation that produced it.
The saved results could not be loaded. No values have been filled in.
Loading the saved experiment…
SAVED DETECTOR OUTPUT
How the different views compared
Results are reported as counts, with timing for the tested images. Missing measurements remain unavailable.
| View | Selected targets matched | Prior negative frames with alerts | Alerts on prior negatives | Mean time per tested image |
|---|
The four older negative controls are separate from the thirteen new originals shown below. An earlier AI review described those controls as having no visible mushrooms. Their alerts are reported separately and still need human inspection. “Not tested” means that comparison was not run.
REVIEW THE ORIGINALS
The images behind the numbers
Choose an image to inspect its original pixels, selected AI reference region and saved results. Broad group names are source hints, not verified identifications.
No images in this round match this filter.
ORIGINAL IMAGE / AI OBSERVATION
Selected original
The original could not be loaded with its recorded dimensions. The reference outline is hidden.
A careful correction is useful data.
Use the lab to mark what you can actually see, including confusing leaves, bark and wood. AI tools can contribute through the separate interface, with their observations kept explicitly unverified.