External testing of models

[jwutsetro]

Dear mazurowski-lab

I am Joris, A PhD student at the VUB Brussels.First of all congratulations on your work, it looks impressive ! In the follow up of the research from Jakub Ceranka ( whom you cited in your Segment any Bone paper) we are also developing new bone segmentation methods on MRI.

Our work is coming to an end before finalising it in a publication. With respect to that, it would be very nice if we could run your data through our model. I assume sharing data is not that easy these days, but when we are ready, we could share our trained models and an inference script.

Potentially we could have a nice collaboration, Extending our datasets by putting them together? A crazy idea lives here to share our data publicly, however, we are currently investigating the steps needed ( ethical approvals, privacy protection etc) to get there.

Looking forward to get in contact with you. If interested, feel free to contact me on [email protected]

[Guhanxue]

Hi, thanks so much for your interest! As you know, due to the sensitive nature of the data and the extensive process involved in the data review and release from the institution, we are currently in the midst of preparing the image data's release as well, so it might be very challenging for us to share data immediately at this moment. What do you think if you could evaluate our model in your data, as we already shared the model and weights? I am also looking forward to future potential collaborations; maybe we can follow up to get more ideas about how to evaluate without violations.