Success at 0.7

[tjdurant]

Hello,

My team and I are trying to use DeepMask and SharpMask for semantic segmentation for medical imaging and in the process of writing up a manuscript, we are having trouble articulating the success at 0.7 metric.

Would someone be able to reach out to us to help clarify how this metric works and why it is being used for selection of 'best model'?

Thank you in advance,
Thomas