Mastering AI Chip Complexity
Explore challenges and solutions in AI chip development
Download eBook
Top Tips to Speed up Orthopedic Image Segmentation
Posted on 30 April 2020 by Celia Butler
The process of carrying out medical image segmentation from MRI and CT can be very time consuming. The quality of the scan data often makes it hard to identify important regions of interest, while final models exported to applications like finite element (FE) simulation or 3D printing may not be robust and require more work before they can be used.
Simpleware ScanIP offers solutions for these problems through common segmentation workflows. We have put together some key methods below that can help accelerate the segmentation of orthopedic data, and to obtain robust 3D models.
1. Use Automated Tools to Segment Image Data
Simpleware recently released a new module, Simpleware AS Ortho, designed to take all the hassle out of segmenting hips and knees. This one-click solution allows you to segment bones and create key landmarks in a few minutes instead of hours!
These tools are currently available when working on images from:
CT of the Pelvic Region: extract the Femurs, Hip (left and right) and Sacrum with a range of medical landmarks generated including: Greater and Less Trochanter, Pubic Tubercles, Anterior and Posterior Superior Iliac Spine, and the Coccyx.
Simpleware AS Ortho hip segmentation tool and output.
PD Weighted MRI of the Knee: extract Femur, Tibia, Fibula, Patella, and selected cartilage. Landmarks include Femur Epicondyles, Tibia Condyles, Tibia Intercondyles.
Simpleware AS Ortho knee segmentation tool and output.
These tools are a time and effort saver for those working with a few to thousands of datasets. Not working with these joints? Contact us to see what we can do for you!
Not working with hips or knees? Contact Us to see what we can do for you.
2. Combine Specialized Tools to Avoid Manual Work
Manual segmentation can be a time drainer when working with complex medical data in 2D and 3D views. In Simpleware software, a number of intuitive tools are available to quickly segment your image data.
For example, if you are looking at a foot, pre-set threshold values can give an estimation of the values needed to segment the bones. Using the Mask Flood Fill tool allows you to create a new mask for each bone which is unconnected. The Split Regions tool can be used to quickly separate neighboring bones.
Use the Split tool in Simpleware ScanIP to separate different bones in the foot.
3. Tidy Up Segmentation with the Help of some Quick Tools
After your initial segmentation with Threshold or Region Growing, you may be left with some tidying up to do. This can be sped up using a range of tools from the Image Processing tab. Use Island Removal to get rid of regions smaller than a specified number of voxels, or Fill Gaps to close up cavities inside objects or at the image boundaries.
Use the Fill Gaps tool in Simpleware ScanIP to close up cavities inside objects.
The Morphological tools are also key for removing thin connections between two bones such as between the Iliac and the Sacrum (using an Open filter), or closing small holes caused by noise or imperfections in the image acquisition process.
4. Enhance Your Segmentation with 3D Editing Tools
Use the 3D Editing tool to identify specific regions of your model and apply filters to them locally. The regions can be selected using a range of tools, for example using a hand drawn Lasso Frustum to deleting connected regions in-between other parts, or a sphere for closing small holes in the Femoral head, or a Cuboid to apply the smoothing filter to a calcification.
5. Solve Problems with Low-Contrast or Poor-Quality Data
CT data presents particular challenges for segmenting muscles using threshold tools due to low-contrast results. To reduce time-consuming manual segmentation, the Interpolation Toolbox in Simpleware ScanIP allows you to combine 3D painting of scan data with a quick interpolation of the results to obtain the profile of the muscle.
Similarly, with MRI data that is low-contrast, you can paint on a selection of slices in each view - a 3D Wrap filter then creates a mask rather than having to paint on every single slice of the image data.
3D Wrap in Simpleware ScanIP allows you to build masks from cross sections.
6. Customize and Automate Workflows
Simpleware software can be augmented for specific workflows with small scripts. Scripting and customization options in the software can be useful to cut down on manual work. For example, if you are working with many similar datasets, scripting can often be utilized to automate repetitive tasks and reduce time of the segmentation.
Use the Simpleware scripting functionality to automate repetitive tasks.
Learn More
We hope these tips provide a good introduction to some of the tools available in Simpleware software to quickly segment orthopedic data. For more details, check out our on-demand webinars on this subject:
Any Questions?
Our technical specialists will be happy to help with problem-solving and taking on unique challenges that you are facing with segmenting your anatomical image data. Want to know about how we can help with your workflow?