Jacquin Antoine 091d7d9eb8 Fix: mask channels below 30 keV in inference and training to prevent misidentification ...

Below ~30 keV the detector signal is dominated by X-ray fluorescence (L-shell)
and artifacts not modelled in training data. This spurious low-energy continuum
caused the model to misidentify Am-241 as Th-232/U-235. Masking channels <30 keV
before inference fixes Am-241 detection from 2% to 99%. Same masking applied in
the synthetic spectrum generator for consistent retraining.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

2026-05-21 21:55:34 +02:00

..

ground_truth

Pipeline complet Radiacode 103 - identification automatique d'isotopes

2026-05-19 12:29:56 +02:00

physics

Fix: CsI(Tl) non-linear response correction + detector calibration overhaul

2026-05-21 17:35:22 +02:00

__init__.py

Pipeline complet Radiacode 103 - identification automatique d'isotopes

2026-05-19 12:29:56 +02:00

config.py

Fix: CsI(Tl) non-linear response correction + detector calibration overhaul

2026-05-21 17:35:22 +02:00

generate_spectra_v2.py

Pipeline complet Radiacode 103 - identification automatique d'isotopes

2026-05-19 12:29:56 +02:00

generate_spectra_v3.py

Background réaliste CsI(Tl) + hybridation mesuré/synthétique + dashboard continuum

2026-05-19 18:14:00 +02:00

generate_spectra.py

Fix: CsI(Tl) non-linear response correction + detector calibration overhaul

2026-05-21 17:35:22 +02:00

generator.py

Fix: mask channels below 30 keV in inference and training to prevent misidentification

2026-05-21 21:55:34 +02:00

spectrum_viewer.py

Pipeline complet Radiacode 103 - identification automatique d'isotopes

2026-05-19 12:29:56 +02:00

visualize_training_data.py

Pipeline complet Radiacode 103 - identification automatique d'isotopes

2026-05-19 12:29:56 +02:00