You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Do you have or do you know where I can get some additional anonymized CT's that are decent mix of study, types, particularly MSK CT's like shoulders, elbows, wrists, hips, knees, ankles, as well as any others.
My current server on the web is a Droplet that is pretty wimpy on Digital Ocean. 2CPU's and 4GB of Ram. I am surprised it even runs with all of the Docker containers I have on there. I might switch over to Amazon shortly.
Also, some guy I've chatted with before might be visiting Vienna for a Conference apparently. He does AI work, so I'd be pretty interested in meeting him in person if he has time. I'd still be interested in seeing how feasible it would be to extract a derived BodyPart from the actual instances using AI and how that compares to what is possible with the tags alone.
I suspect that AI would be better if all of the details could be worked out. Kind of the same story for the Kernels. At least 2 approaches.
The text was updated successfully, but these errors were encountered:
Do you have or do you know where I can get some additional anonymized CT's that are decent mix of study, types, particularly MSK CT's like shoulders, elbows, wrists, hips, knees, ankles, as well as any others.
We would need those for laterality:
"NewName": NewSeriesNameString([Laterality, BodyPartExamined, ImagingPlane, ContrastBolusAgent, SliceThickness, ConvolutionKernel, Projection])
My current server on the web is a Droplet that is pretty wimpy on Digital Ocean. 2CPU's and 4GB of Ram. I am surprised it even runs with all of the Docker containers I have on there. I might switch over to Amazon shortly.
Also, some guy I've chatted with before might be visiting Vienna for a Conference apparently. He does AI work, so I'd be pretty interested in meeting him in person if he has time. I'd still be interested in seeing how feasible it would be to extract a derived BodyPart from the actual instances using AI and how that compares to what is possible with the tags alone.
I suspect that AI would be better if all of the details could be worked out. Kind of the same story for the Kernels. At least 2 approaches.
The text was updated successfully, but these errors were encountered: