MATH/Processes behind Image Classification?

Discussion created by GEOSTUV03 on Feb 6, 2014
Hi there,

I'm doing a MSc dissertation in a "Output Comparison of Image Classification algorithm" with ArcGIS, ERDAS, ENVI and QGIS.

Even-though I use the same algorithm and same settings, the output are different.

Therefore is it possible to get the "processes" or MATH behind how ESRI (and the others) have built their tool, so I can explain why there is a difference ???