Matrix of comparisons between latitude (left), longitude (middle), and diameter between the AI-based database and the reference database. Location is calculated in kilometers and so is a dimensionless unit when divided by avg(D_R, D_db). Values >0 are where the AI-based database has a larger value. Values for diameter do not span the full ±1.0 range (100% difference, e.g., matching a 5 km crater to a 10 km crater) because the matching parameter’s divisor is min(D_R, D_db), while the denominator to compare the data for these graphs is avg(D_R, D_db). The listed means and standard deviations were calculated assuming Gaussian moments, despite nonzero skew and non-Gaussian distributions, so medians and interquartile values are also given. Most of the spikes at ±0 for latitude and longitude have an under-filled bin adjacent, indicating the over/under is an artifact of rounding and aliasing, effectively a Moiré pattern. To interpret these numbers, the first number is the bias relative to R19, while the second number is the 1σ variation.