| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| courses:456:2022:projects:456-2022-28:project-28 [2022/05/20 16:28] – [Technical Approach] mliu90 | courses:456:2022:projects:456-2022-28:project-28 [2022/05/20 16:53] (current) – [Technical Approach] mliu90 |
|---|
| Osteonecrosis of the femoral head (ONFH) refers to the disease that bone cells in the hip joint are dead due to insufficient blood supply to the femoral head. It could lead to failure of the subchondral bone. ONFH affects especially young adults aged between 30 to 50, and the incidence rate of ONFH keeps increasing. | Osteonecrosis of the femoral head (ONFH) refers to the disease that bone cells in the hip joint are dead due to insufficient blood supply to the femoral head. It could lead to failure of the subchondral bone. ONFH affects especially young adults aged between 30 to 50, and the incidence rate of ONFH keeps increasing. |
| |
| {{ :courses:456:2022:projects:456-2022-28:onfh.png?200 |}} | {{ :courses:456:2022:projects:456-2022-28:onfh.png?600 |}} |
| (Image from: https://orthoinfo.aaos.org/en/diseases--conditions/osteonecrosis-of-the-hip) | (Image from: https://orthoinfo.aaos.org/en/diseases--conditions/osteonecrosis-of-the-hip) |
| |
| **MRI Segmentation** | **MRI Segmentation** |
| |
| In this project, the segmentation of MRI volume is firstly realized based on the baseline model --- the nnU-Net. nnU-Net is desired to provide an appropriate segmentation model for the rest of the work in this project. By accomplishing the baseline training, more work will be dedicated to improving the performance of MRI segmentation using 3D CNN models, from the perspectives that: | In this project, the segmentation of MRI volume is firstly realized based on the baseline model --- the nnU-Net. nnU-Net is desired to provide an appropriate segmentation model for the rest of the work in this project. Four networks are designed to develop: 2D U-Net, 3D U-Net, 3D Low-Resolution U-Net and 3D Cascade U-Net. |
| * Loss function: improve the loss function with combinations of Hausdorff distance, dice loss, and negative log-probability. | |
| * Dilation rates: empirically, appropriate dilation rates could improve the model's performance. | {{ :courses:456:2022:projects:456-2022-28:mri_seg.png?400 |}} |
| * Encoder: More advanced visual encoder, for instance, vision transformer, could improve the model's performance. | |
| | Illustration of MRI segmentation result. |
| **CT Segmentation** | **CT Segmentation** |
| |
| For the CT segmentation, the conventional threshold-based method is shown to work on CT femur segmentation, and will be used as the baseline method for this project. The threshold-based algorithm introduced by Krčah et al.[9] will be used as the baseline model for the project. | For the CT segmentation, the conventional threshold-based method is shown to work on CT femur segmentation, and will be used as the baseline method for this project. The threshold-based algorithm introduced by Krčah et al.[9] will be used as the baseline model for the project. |
| |
| | {{ :courses:456:2022:projects:456-2022-28:ct_seg_result.png?600 |}} |
| The improved model for CT segmentation, if necessary, is planned to be another 3D CNN model. The threshold-based segmentation results (with appropriate manual corrections if needed) will then be used as the ground truth annotations for the improved models. Theoretically, the 3D CNN model for CT segmentation can be modified on the basis of the MRI segmentation model. | The improved model for CT segmentation, if necessary, is planned to be another 3D CNN model. The threshold-based segmentation results (with appropriate manual corrections if needed) will then be used as the ground truth annotations for the improved models. Theoretically, the 3D CNN model for CT segmentation can be modified on the basis of the MRI segmentation model. |
| |
| **Method on MRI-CT registration** | **Method on MRI-CT registration** |
| |
| In this project, the manual registration based on segments from MRI and CT will firstly be realized. The segments will be outputted as a segmentation file that 3D Slicer could read. The alignment of segments, as well as the selection of landmarks, will be done manually through the interface. | In this project, a two-stage registration based on segments from MRI and CT are realized. |
| |
| The semi-automatic registration is planned to be realized using Symmetric Normalization from ANTs, an open-source registration toolbox. Using the MRI and CT volume segments (and the initial transformation from manual step), the registration algorithm will then compute the transformation between MRI and CT. Finally, the two segmentation models and the registration model will be integrated together to form an end-to-end system and, if possible, to be integrated with 3D Slicer software. | The semi-automatic (automatic) registration is realized using (1). use ANTs to provide an initial estimation of the transformation between MRI and CT;(2) use MMI-based registration to compute the final transformation. Using the MRI and CT volume segments, the registration algorithm will then compute the transformation between MRI and CT. Finally, the two segmentation models and the registration model will be integrated together to form an end-to-end system and, if possible, to be integrated with 3D Slicer software. |
| | {{ :courses:456:2022:projects:456-2022-28:mri_ct_reg.png?400 |}} |
| ======Dependencies====== | ======Dependencies====== |
| |
