I want to create an automated fracture detection using machine learning from CT 3D volume. What is the best method to do this?

As I am new to MATLAB, I have learnt on the image processing functions but it seems difficult to apply on CT volume (.nii format) as it is in 3D. I have done a lot of literature review suggesting segmentation methods but there is less example to try it on my file. Therefore, what is the methods or guideline to do this?

Answer 1

[yanqi liu]

sir，may be consider some DeepLearning method，such as unet3d

[Wan Faiz]

I researched a lot but cant find related example to apply the deep learning for fracture detection in CT volumes. Have you ever try U-net on related project?

[yanqi liu]

yes，i use it to detect image，such as

[Wan Faiz]

Can you share the methods to do this? Or any reference that I can look into

[yanqi liu]

yes，please use

https://github.com/Andy-zhujunwen/UNET-ZOO

[Wan Faiz]

this python also can be applied to matlab?

[yanqi liu]

yes,use it as console app or http api to run in python and get result in matlab

[Wan Faiz]

Can i know the graphic memory requirement for this 3d architecture