Tanguy Duval

Imaging of brain tumors for the neurosurgeon • Development of medical imaging for brain tumors I proposed and developed the entire acquisition and processing pipeline: I designed the multimodal (spectroscopy, spin-diffusion, spin-relaxation, and spin magnetization-transfer effect) imaging protocol on the instruments and tested it on phantoms for calibration. I developed a post-processing environment with Docker-encapsulated modules (artifacts corrections, noise reduction, distortions correction based on the mapping of the magnetic field, and bio-physical modeling for interpretation of images and detection of infiltrative tumourous cells). I finally proposed and wrote operating procedures for data organization and processing modules creation. I used this framework to extract a big database from the hospital and trained a custom generative deep learning model for patients with Glioblastoma. • Consultant on the Neurostars project – Franco-Dutch consortium (with private companies), financed by the H2020 (Eurostars program). The project was to develop visualization (desktop and virtual reality) of imaging data for the neurosurgeon. I was the only imaging expert, so I was in charge of the specifications, the norms, and the control of the deliverables (schedule and testing). • Fundamental research on tumor cell detection I proposed a study (received funding, 50k), for the imaging of brain tumors with advanced MR protocol. Biopsy extraction and characterization (automatic segmentation and counting of stained tumor cells) was proposed as ground truth.