You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks @scuniff and agreed that the width and the height should be casted to double to handle all scenarios. In practice, all the FV1000 samples in the OME curated data repository have even values for their width and height so this rounding issue never came up as part of the testing.
Feel free to open a Pull Request with your proposed changes and we can check it using the nightly CI builds. As a bonus, if you happen to have a sample FV1000 dataset allowing to test this condition, feel free to upload it to https://zenodo.org/
bioformats/components/formats-gpl/src/loci/formats/in/FV1000Reader.java
Lines 1348 to 1349 in 0a895ae
Since width and height are ints, the equation is done with Integer accuracy and not Double accuracy.
To get Double accuracy from the equations, a double cast is needed
Maybe
double rx = (double) width / 2;
double ry = shapeType == CIRCLE ? rx : (double) height / 2;
The text was updated successfully, but these errors were encountered: