monocongo/segment_maskrcnn

Add dataset support for annotations JSON from VIA

Closed this issue · 1 comments

In order to use segmentation annotations produced by the VGG Image Automator tool, we'll extend the functionality of the maskrcnn.dataset module to allow for using an annotations JSON file produced by VIA instead of individual mask files.

This appears to have been resolved by PR #13 (VIA annotations support was validated using the notebook included in that PR).