Semantic Segmentation Choroid Plexus
I am starting on a new project to segment Choroid Plexus. Choroid Plexus is a very small organ inside the brain. Below is the problem and objectives.
1) Choroid plexus shows large inter-patient variation in term of shapes and sizes.
2) Choroid Plexus segmentation constitutes a difficult task due to low tissue contrast and large
variability in organ shape and size.
3) Choroid Plexus boundary delineation is a difficult task due to shape-variability and poor tissue
1) 3D CNN which could learn the geometry of choroid plexus.
2) Train the network on smaller 3D regions centered around the choroid plexus instead of the
whole T1 Brain MRI Image.
3) Reduce the search space significantly by ruling out parts of scans that are very unlikely to
contain choroid plexus, since choroid plexus is expected to be inside the ventricles.
4) Segmentation frameworks that rely on additional preceding object localization models to
simplify the task into separate localization and subsequent segmentation steps.
5) The model should give more importance to the choroid plexus channels/feature maps and
also the location of the choroid plexus.
6) An architecture for more non-rigid anatomical organs such as choroid plexus, because their
topologies varies much, meaning there is a need to learn many low-resolution high-level
It would be very helpful to know models, research papers which partly/fully addresses the above objectives. Thank you so much for your help. I am looking forward to everyone inputs