Closed hackermd closed 8 years ago
The current approach for saving and retrieving objects is based on outline coordinates of objects.
skimage.measure.find_contours()
cv2.fillPoly()
It needs to be tested whether this approach works reliably for different types of objects...
The whole approach has changed in the meanwhile. Outlines are stored as polygons in the database and the label images are reconstructed from the polygon coordinates.
The current approach for saving and retrieving objects is based on outline coordinates of objects.
skimage.measure.find_contours()
cv2.fillPoly()
It needs to be tested whether this approach works reliably for different types of objects...