Development metrics

[FabianIsensee]

Computation of metrics, SD tolerances, sampling of realistic segmentations from instance annotations

Description

• metric computation (dice and SD)
• documentation for how to use the code to evaluate segmentations (see evaluation/readme)
• code for sampling realistic segmentations from instance annotations
• code for computing tolerances (needed for SD) based on these sampled segmentations
• code for computing inter-rater disagreement on sampled segmentations

(- updated PR template to also list setup.py version update in checklist)

Checklist

• [x] Merged latest master
• [x] Updated version number in README.md
• [ ] Ran annotation.import to completion -> cannot do this
• [x] Added changes to changelog.md