Semi-automatic micro-CT segmentation of the midfoot using calibrated thresholds.

Melissa R RequistYantarat SripanichAndrew C PetersonHazibullah Waizy Alexej Barg Amy L Lenz

Published in: International journal of computer assisted radiology and surgery (2021)

We developed a semi-automatic micro-CT segmentation method based on calibrated thresholds. This method was designed specifically for use in bones with high rates of curvature and low cortical bone density, such as the cuneiforms, where traditional threshold-based segmentation is more challenging. Our method shows improvement over manual segmentation and was highly reliable, making it appropriate for use in quantitative micro-CT analysis.

Keyphrases

deep learning
convolutional neural network
image quality
dual energy
computed tomography
contrast enhanced
positron emission tomography
machine learning
magnetic resonance imaging
neural network
data analysis