Evaluating input graphical parameters for the training phase of designing an intelligent emotional vector deducing engine

摘要

This paper addresses graphical input evaluation problem from IAPS photos for the training phase of designing an intelligent emotional vector deducing engine, which would simplify the affective computer interfaces. The experiment results analyzed the contribution of input parameters for the training phase. This paper further addresses the affect of increasing the 4 added parameters, sums of edges, corners, regional maximums and regional minimums of affective photographs which proved to increase the training accuracy.