Semi Supervised Segmentation for 3D cardiac data
Levreging previus work
If you are intrested in one of the projects, BSc, MSc thesis (Technion/BGU) or academic collaboration, contact me by mail! If you have interesting problem/data and want to offer projects to the students in the Technion, we can explore them as well.
Levreging previus work
By adverserial attack
Generate new insights about how and what to mix
Formulate the segmentation task as regression only
Investigate the biasses of self annotation and how to combat noisy labels in segmentation tasks
In this project we will exemine changing NN architecture to be more robust / less computationaly heavy, using classifier free diffusion models