Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Learning implicit deep tissue structures for Near infrared optical tomography
It is a challenge to perform image reconstruction for near infrared optical tomography in real time based on conventional model-based approaches. Machine learning methods using implicit shapes have a great potential to improve NIROT image reconstruction.
Keywords: Deep learning, image reconstruction, medical imaging, diffuse optics, machine learning, time-of-flight imaging
Cancer is a highly diverse disease with a large variability of effectiveness in therapies from patient to patient. Therefore, one of the main aims of current research is to find, investigate and measure biomarkers that indicate how to personalize the treatment of cancer. Tumor oxygenation is one of the most important biomarkers. To image hidden tumor non-invasively, we developed a single photon avalanche diode (SPAD) camera -based imaging system Pioneer. The big data generated with the SPAD camera on the one hand improves the image quality, on the other hand, it is a challenge to perform image reconstruction in real time based on conventional model-based approaches. Implicit representations of 3D shapes have gained popularity in simultaneous localization and mapping (SLAM). The use of implicit shapes has a great potential to improve NIROT image reconstruction.
The project involves 1) learning the basic physics of NIROT, 2) development of learning approaches, 3) validation of the method in simulation 4) test on measured data.
We are looking for self-motivated students with background in computer science, biomedical engineering, electrical engineering, mechanical engineering or related subjects. Python / MATLAB is required for this project. Knowledge of SLAM / CV/ DL is a plus.
Cancer is a highly diverse disease with a large variability of effectiveness in therapies from patient to patient. Therefore, one of the main aims of current research is to find, investigate and measure biomarkers that indicate how to personalize the treatment of cancer. Tumor oxygenation is one of the most important biomarkers. To image hidden tumor non-invasively, we developed a single photon avalanche diode (SPAD) camera -based imaging system Pioneer. The big data generated with the SPAD camera on the one hand improves the image quality, on the other hand, it is a challenge to perform image reconstruction in real time based on conventional model-based approaches. Implicit representations of 3D shapes have gained popularity in simultaneous localization and mapping (SLAM). The use of implicit shapes has a great potential to improve NIROT image reconstruction.
The project involves 1) learning the basic physics of NIROT, 2) development of learning approaches, 3) validation of the method in simulation 4) test on measured data.
We are looking for self-motivated students with background in computer science, biomedical engineering, electrical engineering, mechanical engineering or related subjects. Python / MATLAB is required for this project. Knowledge of SLAM / CV/ DL is a plus.
The aim of this project is to develop new data-driven methods for time domain near infrared optical tomography (TD NIROT) to reveal tumor shapes and locations hidden in tissue.
The aim of this project is to develop new data-driven methods for time domain near infrared optical tomography (TD NIROT) to reveal tumor shapes and locations hidden in tissue.