Posture-invariant ECG recognition with posture detection

Recently Electrocardiogram (ECG) has been proposed as a biometric modality which offers liveliness detection. The fact that ECG is a vital signal makes it challenging to work with as it is affected by physical and psychological changes. In realistic applications, this type of biometrics still needs to be verified in conditions related to the practical use. In real life our body posture changes frequently, therefore in the context of a biometric system our body posture may be different in enrolment and verification which can potentially decrease the performance of the system. In this paper we first investigate the effect of the body posture on the accuracy of ECG biometric systems. Second, a new method is presented that is able to clearly distinguish the ECG signal of different postures of an individual. Finally, we propose a posture-detection verification system in order to mitigate the effect of body posture by first detecting the posture of a subject and then identifying it.