How do we implement a benchmark?

[dbuscombe-usgs]

It seems prudent to document progress and assess CoastSeg (and CoastSat) workflows against a ground-truth shoreline dataset, supported by a standardized workflow. In other words, a benchmark that consists of (at least)

• published shoreline data from survey locations with long time-series of shoreline data from GPS or other survey equipment
• a set of ROIs for those regions
• a workflow for comparing the shorelines obtained from CoastSeg with the published shoreline data

[dbuscombe-usgs]

See also this parallel discussion: #43 (reply in thread)

Perhaps it makes most sense to participate in this benchmarking exercise rather than reinvent the wheel.