PET scientist · Inverse problems · Efficient deep learning
Postdoctoral researcher at LaTIM (INSERM UMR 1101, Université de Bretagne Occidentale, Brest) 🔬. PhD on deep learning for PET reconstruction and data correction (thesis).
| Project | Topic | Paper | Code |
|---|---|---|---|
continuous-3d-latent-diffusion |
Continuous 3D latent diffusion for CT/MRI generation and reconstruction | Preprint (under review) | Code |
ddconv-prc |
Dual-input dynamic convolution for positron range correction | IEEE TRPMS 2025 · arXiv | Code |
Direct-Three-Gamma-PET |
Direct three-gamma PET image reconstruction | IEEE TRPMS 2025 · arXiv | Code |
Particles_Tracking_GAN |
GAN-based fast simulation of F-18 positron paths | ISBI 2024 · arXiv | Code |
| OCT toolbox | Quantification of experimental uveitis in 3D murine retinal OCT | Biomed. Opt. Express 2023 | Code |
Full list of publications: mellak.github.io.