X-ray in-line phase contrast micro-tomography is quickly gaining acceptance as a standard coherent X-ray imaging technique. This is especially true in topics such as paleontology and materials science, since adapted algorithms exist for homogeneous objects [1, 2], which is often the (nearly) the case in these disciplines. In the life sciences, however, we almost exclusively deal with heterogeneous samples, such as in the study of bone and soft tissue in the knee joint, the bone/cartilage interface in osteoarthritis or bone healing. While direct phase retrieval methods for heterogeneous samples have recently been developed [3], these require images at several distances to reconstruct the phase. This adds to the experimental complexity and increase in use of expensive beam time.
The aim of this project is to propose, implement and apply iterative tomographic reconstruction algorithms for in-line phase tomography. The iteration in the object domain would allow the introduction of more sophisticated priors, such as a discrete map of relationship between absorption and phase; and priors such as a penalty on the L1 or TV-norm, which are not appropriate to apply in the Radon domain in general. For the iterative tomographic reconstruction part the forward problem has to be modeled, which we propose should be done on the VIP platform. This modality is not yet available on VIP.
The new reconstruction algorithms will be applied in the imaging of mouse knee joints as models in the study of osteoarthritis. Ideally, soft tissue, bone and vascularization should be imaged simultaneously. Phase contrast alone does not yield sufficient contrast for intra-bone vascular imaging. Therefore imaging will be performed using contrast agents, creating strongly heterogeneous samples, requiring the use of the most recent reconstruction algorithms.
The candidate should have a strong background in applied mathematics, biomedical engineering, and/or computer science. Experience in X-ray imaging in general and phase contrast in particular is desirable. Proven communication and scientific writing skills in English is necessary.
Send your applications and inquiries to Max Langer (max.langer@esrf.fr)
Keywords: X-ray phase tomography, tomographic reconstruction, optimization, mathematical modeling, osteoarthritis, small animal imaging, matlab, simulation
References
[1] Langer et al., IEEE TIP 2010
[2] Paganin et al., J Microsc 2002
[3] Langer et al., Opt Lett 2012