Keywords
Deep learning, convolutional neural network, spectral CT.
Project
X-ray “color” or “spectral” computed tomography (CT) is a new imaging modality that is raising increasing interest in radiology. Thanks to the emergence of new detectors that can discriminate X-ray photons depending on their energy [1], it is possible to reconstruct the constituents of the human body such as bone, water, fat or concentration in contrast agents. Although recent works have shown the feasibility of spectral CT systems, there are still many open questions such as the best way to decompose and reconstruct images of the object into a material basis. We recently proposed a model-based material decomposition method [2] that requires the knowledge of source and detector response functions. However, these are generally unknown or difficult to model.
Work Plan
The goal of the internship is to perform material decomposition using deep learning in order to circumvent modelling the source and detector response functions. In particular, we will investigate various deep learning architectures [3] and compare them to model-based approaches. The successful candidate will address the following points:
Generation of training data sets using in-house Matlab toolbox
Implementation of deep learning methods using the TensorFlow library in Python
Comparison with existing (model-based) methods using simulated and experimental data
Collaboration with CPPM (Marseille) and UCL (CMIC, London, UK)
Salary
550€ net monthly
Skills
The student must have a strong background image processing and deep learning. Knowledge in medical imaging and radiation physics is appreciated but not required. Programming skills: Matlab, Python.
How to apply?
Send both your CV and academic records to Juan FPJ Abascal (juan.abascal@creatis.insa-lyon.fr) and Nicolas Ducros (nicolas.ducros@creatis.insa-lyon.fr)
Reference
[1] K. Taguchi et al, “Vision 20/20: Single photon counting x-ray detectors in medical imaging,” Medical Physics, 40, 100901, 2013.
[2] N. Ducros et al. “Regularization of nonlinear decomposition of spectral X-ray projection images” Medical Physics, 44, 9, e174-e187, 2017.
[3] Y. LeCun et al. “Deep learning”, Nature 521, 436–444, 2015.