List view
1. [ ] Data preparation 1. [ ] Data conversion for voxel2mesh 1. [x] LIDC 2. [ ] LUNGx 2. [ ] Outcome 1. [ ] LIDC 2. [ ] LUNGx 2. [ ] Nodule classification 1. [ ] Build a classification using voxel2mesh features with LIDC train&val set and evaluate it with LIDC72 2. [ ] Transfer learning using LUNGx calibration and evaluate it with LUNGx test 3. [ ] Intention-to-submit by 23:59, Pacific Time, 15th February 2022 4. [ ] Full manuscript submission by 23:59, Pacific Time, 1 March 2022
Overdue by 3 year(s)•Due by March 1, 20221. [ ] Nodule Detection 1. [ ] Create an interface to connect existing detectors including [pylidc](https://pylidc.github.io/) and lungx database 2. [ ] Nodule Segmentation 1. [ ] Angle Preserve Deformation - add a loss component 1. [ ] CPU computation for testing 2. [ ] GPU computation using [PyTorch3D](https://pytorch3d.org/) 2. [ ] Improve code quality - [MeshRCNN](https://github.com/facebookresearch/meshrcnn) 3. [ ] Spiculation Quantification 1. [ ] Recursive searching algorithm in matlab -> ? 4. [ ] Nodule Classification 1. [ ] A simple classifier using two FC layer 2. [ ] ?
No due date