Domain, Specialties : Physique nucléaire
Keywords:
Signal processing, Machine learning, Nuclear physics
Research Unit : DEDIP / LILAS
Summary
This internship aims at developing deep learning methods for spectral unmixing in gamm-ray spectrometry.
Full description
DESCRIPTION AND PROBLEM
Gamma-ray spectrometry is a classical technique used to identify and quantify radionuclides in a wide range of nuclear applications (environment, industry, metrology, illicit trafficking, etc.).
An observed spectrum is made up of M channels corresponding to energy intervals deposited in a photon-sensitive detector (scintillator or semiconductor). The count in each channel represents the number of events in a given energy interval. For each radionuclide, there is a characteristic spectrum called the spectral signature, corresponding to the detector response.
Let X be a matrix, each column of which is the normalized spectral signature of each radionuclide. The observed spectrum follows the Poisson distribution of Xa:
Location
Saclay
Internship conditions
- Internship duration: 6 months
- Level of study: Bac+5
- Training: Master 2
- Continuation in PhD thesis: Yes
- Application deadline: 2 juin 2024
Experimental skills
Language : English
Useful methods and technics:
Deep learning, spectral unmixing
Computer languages and software:
Python, pytorch