Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Deep learning: Super-resolution HR-pQCT imaging to investigate mechanoregulation in patients
To investigate bone mechanoregulation in diabetic patients, a deep learning-based method for high-resolution reconstruction of trabecular microstructures from low-resolution HR-pQCT scans using GAN-CIRCLE will be developed
Bone fragility is dependent not only on bone mass but also on microstructure and the intrinsic material properties of the tissue. Based on recent advent of HR-pQCT, we have developed imaging and computational methods allowing monitoring of local changes in the bone microstructure over time and calculating local mechanical loading. It has been shown that high spatial resolution is required to study cellular behaviour – in form of bone remodelling sites – and calculate the corresponding mechanical loading. However, in clinical practice resolution of HR-pQCT images is limited (61 micron). In order to surpass this barrier and push our computational tools from “bench” to “bedside”– modern image processing and computer vision methods are required.
Bone fragility is dependent not only on bone mass but also on microstructure and the intrinsic material properties of the tissue. Based on recent advent of HR-pQCT, we have developed imaging and computational methods allowing monitoring of local changes in the bone microstructure over time and calculating local mechanical loading. It has been shown that high spatial resolution is required to study cellular behaviour – in form of bone remodelling sites – and calculate the corresponding mechanical loading. However, in clinical practice resolution of HR-pQCT images is limited (61 micron). In order to surpass this barrier and push our computational tools from “bench” to “bedside”– modern image processing and computer vision methods are required.
Goal of this project is to develop a deep learning-based method for high-resolution reconstruction of trabecular microstructures from low-resolution HR-pQCT scans using CNNs.
Goal of this project is to develop a deep learning-based method for high-resolution reconstruction of trabecular microstructures from low-resolution HR-pQCT scans using CNNs.
Due to corona, we may also host remote projects (based on pre-existing clinical, or rodent in-vivo data) which do not require physical presence but a stable internet connection.
Please do not hesitate to contact me with an informal e-mail (matthias.walle@hest.ethz.ch) so we can work out a project tailored to your interests!
International students (within the EU, due to COVID) are welcome to apply!
Due to corona, we may also host remote projects (based on pre-existing clinical, or rodent in-vivo data) which do not require physical presence but a stable internet connection.
Please do not hesitate to contact me with an informal e-mail (matthias.walle@hest.ethz.ch) so we can work out a project tailored to your interests!
International students (within the EU, due to COVID) are welcome to apply!