Today, Armand Delbos presented his results on deep learning for microvascular imaging at Resolve Stroke. With Arthur Chavignon and Maxence Reberol, Armand has developed a brilliant implementation of 4D U-Net pipeline for ultrasound imaging. Convolutional Neural Networks (CNNs), and in particular U-Nets, have been extensively used in medical imaging due to their ability to capture fine and complex feature representations through a series of convolutional and pooling layers. In traditional transcranial imaging, the strong absorption of ultrasound waves passing through the skull bone distorts the images, making anatomical features difficult to decode. By implementing a 4D U-Net pipeline, Armand was able to leverage a complete 3D+t dataset to improve imaging performances. Stay tuned for the results. Bravo, Armand Delbos 👏👏
🤩 Very proud of this team!
Bravo Armand !
Congrats Armand Delbos !
Félicitations Armand Delbos !
Congratulations Armand Delbos !
Let's make this thing work
4moLong live the networks who use all available dimensions in the input dataset (and their patient trainers)