For a segmentation project with sonographic images we would like to create the ground truth using MITK. We have images from different vendors and in different formats, i.e. "normal" 2D stills and "movies" 2D+t.
I noticed the following while using the dicom data with MITK:
- The Canon images (stills) cannot be read (see screenschot attached)
- Philipps movies and Canon movies can be read without error. However, they are displayed a) turquoise and b) as a 3D volume (which sort of is correct)
- Phillips stills can be read without error, but are also colored in turquoise (although the colormap is set to grayscale, changing the colormap does not have an impact). Adjusting the levelwindowing is not easy / hardly possible wrt finding optimal thresholds for segmentation
- Multilabel segmentation is for neither vendor or format available (series does not show up in drop down for "Patient Image"
Attached are the various images and formats, hopefully named in a self-explanatory way.
Tested using MITK v2018.04.2 for Mac (newest binary from the webpage)