Ok, I think you are talking about target shape.
Detector designers do not program to explicitly detect target shape. Detector designers are typically looking for a target to exhibit certain characteristics when the transmitted radio wave emanating from the coil hits the target. The radio wave when it interacts with a metallic target induces a current in that target. That current creates a magnetic field around the target which is then detected by the receive portion of the coil. Detector designers program the detector to sense various characteristics of the target including its conductivity (ability to pass a current through it, which is determined by the material makeup, e.g., silver is a high conductor and mass of the target (i.e., massive low conductive material can look highly conductive to a metal detector), its inductance (how stable the current remains as the transmitting wave intensity is changed - this is frequency dependent and is why certain materials are more detectable at either higher or lower frequencies, it is also shape dependent with round objects like coins exhibiting the most stable magnetic fields indicating high inductance) and finally, they are looking for the natural magnetic response from the target (ferrous targets have a high magnetic response, while non ferrous targets have a low or non-existent natural magnetic response). All this information is derived from the difference in phase angle between the transmitted signal and the received signal and is how the target ID is developed. Since the phase angle can only differ by a maximum of 180 degrees you can get wraparound effects where ferrous signals wrap around and get picked up as very high falsing signals.
Even though, in general, iron objects of any shape exhibit low inductance, since inductance increases with frequency and since round targets (ferrous and non-ferrous) in general exhibit high inductance, some large, round ferrous targets can be mischaracterized by the detector as non-ferrous due to the high inductance associated with their shape and high conductance associated with their mass. So in effect, metal detectors infer a round target shape by the uniform, symmetric characteristics (i.e., inductance) of the field they are sensing, which are likely coin like or ring like therefore they are biased to produce strong signals or stable ID's for the detectorist. That is also why round, but broken objects like hoop earrings may not result in a stable ID or strong signal because the field becomes distorted due to the broken magnetic and electric circuit.
Trying to simplify a very complex subject here so I may have messed up some of the analogies here but I hope it gets the point across. In a nutshell, detectors are looking for a nice symmetric induced magnetic field which implies a round object, like a coin or ring and they tend to reject (through discrimination circuits) targets that have a high natural magnetic response, low inductance, and low conductance (typical of ferrous targets). But since all of these things tend to overlap at points, the detector can get fooled and the result is falsing or inaccurate target IDs. It is actually amazing detectors do as well as they do to give the user a pretty good ID of what is in the ground most of the time given the number of variables (target shapes, ground effects, mixed metals, target orientation, depth, multiple targets under the coil, etc.) they have to attempt to compensate for.
HTH
