The aim of the thesis is to design advanced unmixing and classification algorithms taking into account all the spectral information provided by X-ray hyperspectral detectors to improve the performances of the global systems either for medical imaging, scientific instrumentation and control for security.
As part of its X-ray (RX) imaging developments, LETI is studying the contribution of new CdTe-based hyperspectral RX detectors combined with advanced processing methods. The main applications are medical imaging, scientific instrumentation and control for security. The laboratory works in particular on X-ray detection systems of illicit substances such as explosive materials in air transport.Current data processing methods for discriminating materials or tissues analyzed from measurements are derived from techniques used with dual energy detectors.The aim of the thesis is to design advanced unmixing and classification algorithms taking into account all the spectral information provided by the detectors to improve the performances of the systems in terms of false alarm rate and good detection rate. The challenge is to demonstrate that these detectors and their associated data processing make it possible to achieve performances specified by equipment certification authorities. The proposed methods might be inspired by spectral unmixing and classification techniques widely developed in the context of hyperspectral imaging for Earth observation.The candidate must be specialized in signal processing and show interest in physics and instrumentation.
CEA Grenoble FRANCE
PAULUS Caroline / Phone number: 0438782563 / Email: email@example.com
UNIVERSITY / GRADUATE SCHOOL
Grenoble INP : Electronique, Electrotechnique, Automatique, Traitement du Signal (EEATS)
Start date on 01-10-2018
(c) GdR 720 ISIS - CNRS - 2011-2018.