Closed yoonchoi-neuro closed 1 month ago
Dear Yoonchoi-neuro,
Thank you for your interest in testing our tool and providing us some valuable feedback, we will work on providing a more descriptive guideline to use the software with Docker.
I will use your last message to give some examples. The previous missing element was the
sudo docker run -v /hdd2/ABIDE/image_data_T1/:/data/in -v /hdd2/ABIDE/test/:/data/out kilianhett/chp_seg:1.0.0 --sequence_type T1 --name_pattern ABIDE
/hdd2/ABIDE/image_data_T1 has to be a folder that contains all images (the software will not look into subfolders)
/hdd2/ABIDE/test/ is the target folder where all the output will be stored.
--sequence_type The current version of our software has been trained separately using T1, T2, or FLAIR sequence. As you want to run the segmentation software on T1-weighted images, the option you put is correct.
--name_pattern This will filter out scans based on their filenames, playing as follows for the option "ABIDE" = ABIDE (please do not add * as the regular expression is specified in the code). Also, note that this option is optional, without this option the software will be run on all files in the folder /hdd2/ABIDE/image_data_T1 if the folder contains other types of files (such as .json), I recommend you use ".nii".
Thank you, Kilian
Dear Kilian Thanks for the helpful descriptions. Could you elaborate also on the text output of the software? The last line returns “Total volume = xxxx voxels”. When I compare this to the output of 'fslstats -V' (number of non-zero voxels in mask / volume of mask in mm3) I see that the values do not align. So what is the last line of the text output referring to? Thanks!
Hello Tobias,
The output was a remaining code to monitor the segmentation. It has no value as it corresponds to the number of voxels of the mask before back-transformation to the native input scan space.
Also, I wanted to let you know that we just released a software upgrade version. We corrected this confusing point and extended the estimation of the volume for the right and left ChP in addition to the total volume. Finally, the software now generates automatically volumetric reports in CSV format.
Best, Kilian
Excellent - thanks, Kilian! I tested the new version - works like a charm.
I'm happy to hear it. I pushed small fixes to make the code more robust, don't hesitate to pull the docker again if you want to have the last version.
I am closing this thread.
Kind regards, Kilian
Dear github master,
First of all, thank you for sharing such a useful tool. I followed all the instructions to apply the tool to images, but it is a bit hard to use the command for people not familiar with docker.
Can you please give any specific command example?
Thank you.