Crossover Error Rate (CER)

The crossover error rate, also called the equal error rate, comes from plotting a biometric system’s false acceptance rate and false rejection rate against its matching threshold. Tightening the threshold makes an impostor less likely to be accepted and a genuine user more likely to be refused, so the two curves move in opposite directions and cross at one point. The rate at that crossing is the CER. Study sources commonly use it as a single figure for comparing biometric devices, where a lower CER indicates better accuracy at that point.

The limit is in the last three words. A deployment runs at the threshold its risk needs, not necessarily at the crossover: a high-security door accepts more false rejections to push false acceptance down. The better purchase is the device that performs better at the chosen operating point, which need not be the one with the lower CER.

Exam relevance: questions in this area tend to give error rates for several devices and ask which is most accurate, where the lowest CER is the expected comparison. A scenario that adds a strict security requirement is likely to turn on the operating threshold instead, so candidates are expected to know which way each rate moves as the threshold changes.