Enhancing TEAM Engine with Computer Vision to evaluate correct portrayal of features

[ghobona]

It is not always possible to evaluate whether Client Applications are correctly interpreting SLD and SE.
What if we could enhance TEAM Engine with a Deep Learning Computer Vision capability that evaluates a screenshot from a Client Application and reports on how well the client has portrayed the symbol for a feature.

Note that the Technical report from the DGIWG Portrayal Technical Panel testing of SLD (1.1.0) for OGC Discussion Paper http://docs.opengeospatial.org/dp/17-059/17-059.html[(17-059)] identified differences between QGIS and GeoServer portrayals of the same features from the same SLD/SE document.