SystemOnline
Volcanic / IMP
IMP CANDIDATE
24.540°N 8.067°E · NEARSIDE · NORTHERN HEMISPHERE
- Predicted areaPredicted area
- 2,748 m²
- Tile shareTile share
- 4.2 %
- Mean softmaxMean softmax
- 0.92
uncalibrated
The NASA-IBM LFM segmentation model predicts an irregular mare patch candidate covering 2,748 m² of this 256 m NAC tile. It is a model segmentation of a volcanic feature candidate, not an active volcano. This is a training sample: the model saw this area during training, so the prediction is not independent evidence of model performance.
Why MOONATLAS selected this
IMP candidate
Largest predicted IMP area of its observation group. Product curation over model outputs, not a scientific ranking.
Dataset split
The model saw this area during training, so this prediction is not independent evidence of performance. The observation and the model output are equally real.
M1245967286RE · 256 m tile · 1.00 m/px · masks at native resolution
A segmentation of a volcanic feature candidate, never an active volcano. The mean softmax score is uncalibrated, so it ranks pixels rather than stating a probability.
Model outputs
- PREDICTED AREAAI prediction · NASA-IBM LFM
- 2,748 m²
- IoU VS REFERENCEMOONATLAS derived metric
- 0.648
Provenance
- Model
- NASA-IBM Lunar Foundation Model
- Task
- Irregular mare patch segmentation on LROC NAC imagery
- Checkpoint
- ni_lfm_ps8_frozen_s44.ckpt
- Dataset
- sombench-imp-segmentation
- Split
- Training sample
- Source
- M1245967286RE.ech.cog__target_2470__idx_2470_p0_img.tif
- Processing
- science-0.3.1
- Data origin
- Real model output
Scientific details
- Model id
- lfm-imp-segmentation
- Model revision
- 10a09ffae35e
- Reference
- hargitai-2025
- Object id
- imp-m1245967286re-t2470-i2470
- Catalogued
- 2026-09-17
Values are model output as produced, or MOONATLAS derived metrics computed from it; reference annotations are shown separately and never mixed into predictions. How MOONATLAS works.