Testing GUI

The testing GUI should allow the user to specify an image and a text file like with training. The user would be able to specify multiple pipelines (algorithm configurations). It should be possible to save and load pipelines and groups of pipelines. The testing system should automatically re-run training (or parts of training) when algorithms have been updated or added. The appropriate sequence would be this:

1. The user specifies multiple pipelines
2. The user specifies the image and text file
3. The user is allowed to correct the image segmentation
4. All of the algorithms combinations are run, separately, with output about progress (and preferably also information about which features they detected etc, displayed visually if appropriate).
5. The program displays an accuracy comparison for each pipeline