After the outbreak of severe acute respiratory syndrome (SARS) in 2003, many international airport quarantine stations conducted fever-based screening to identify infected passengers using infrared thermography for preventing global pandemics. Due to environmental factors affecting measurement of facial skin temperature with thermography, some previous studies revealed the limits of authenticity in detecting infectious symptoms. In order to implement more strict entry screening in the epidemic seasons of emerging infectious diseases, we developed an infection screening system for airport quarantines using multi-parameter vital signs. This system can automatically detect infected individuals within several tens of seconds by a neural-network-based discriminant function using measured vital signs, i.e., heart rate obtained by a reflective photo sensor, respiration rate determined by a 10-GHz non-contact respiration radar, and the ear temperature monitored by a thermography. In this paper, to reduce the environmental effects on thermography measurement, we adopted the ear temperature as a new screening indicator instead of facial skin. We tested the system on 13 influenza patients and 33 normal subjects. The sensitivity of the infection screening system in detecting influenza were 92.3%, which was higher than the sensitivity reported in our previous paper (88.0%) with average facial skin temperature.