I am working on a project related to automating the procedure of manually segmenting some bones in CT scans and hopefully if everything goes alright in this stage, move on to do something more with them - like bone reconstruction etc.

I have been doing extensive research regarding this - and CNNs were something in my target as a ML method that could be used here. Emphasis is more on using Deep learning for this project.

So, what I have - the data: CT scans of chest/shoulder and for each of the CT scan, I have 4-6 STL files of the individual bone fragments or segments located in the shoulder or near shoulder region. I am a tad uncertain as to how to use those individual STL files.

Target: To label/classify/identify these fragments in the CT scan - automate it.

My MOA (Method of Approach) or what I understand - I believe it is object (bone fragment being the object) detection and feature (of those bone pieces that I need to lock on in the CT-scan) extraction using CNNs. I am looking at Mask R-CNN etc, use a pre-trained CNN for this.

But I am not entirely sure if my understanding is correct. This is my first time with this stuff, but hoping to learn more. CT-scans are in nifti format.

I could provide more info if required, would gladly appreciate any insight or help or advice with what could be the way forward and if I am thinking along the correct lines.

Thank you.

  • $\begingroup$ I am also working on this project, can you please share your dataset $\endgroup$
    – kkldc
    Commented Aug 23, 2023 at 18:54


You must log in to answer this question.