It is a considerable task to label a large amount of images for the purpose of training deep machine learning algorithms efficently.
Is there software that you would recommend for this purpose, especially freeware;
a) for dichotomous classification
b) for multivariable classification of images?
Kind regards,
JB
I can recommend two tools that were very useful for annotations in my works:
1) Sloth (http://sloth.readthedocs.io/en/latest/);
2) labelImg (https://github.com/tzutalin/labelImg)
Both are easy to install/use and can be customized for different formats of annotations.
My colleague provided an open labeling GUI for image quality assessment for MRI: https://sites.google.com/site/kspaceastronauts/iqa/labelingplatform
DICOM is supported.
Digits provides very easy platform for arranging your data for CNN, Caffe and tensorflow platform.
You Just make folder of different classes and paste your images in each folder. The Digits considers on folder name as label for all images.
https://github.com/NVIDIA/DIGITS
Dear colleagues,
thank you for your very helpful feedback!
Kind regards from Cologne! JB
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training