Probability of detection (PoD) models take into consideration one or multiple flaw parameters, say the length, maximum depth, or surface area, and a single flaw signal. However, due to the correlation between the response signals, it might be necessary to consider multiple response signals to know about the flaw. Hence, in this work, we demonstrate the possibility of including multiple correlated flaw signals (features) towards the construction of a PoD curve. The flaw features considered are the peaks of the 3 components of the magnetic flux density i.e. B_x, B_y, and B_z. This is a fully model assisted (FMA) study in which a non-magnetic steel plate, that contains a narrow opening flaw, is inspected by an eddy current probe that induces spatially uniform fields in the conductor in the region of interest. The analysis was performed using the semi-analytical boundary element method.