An Introduction to Multimodal Biometric System: An Overview
Mamta Ahlawat1, Dr. Chander Kant2
1Research Scholar, Department of Computer Science and Application, K.U., Kurukshetra, INDIA
mamtaahlawat999@gmail.com
2Asst.
Professor, Department of Computer Science and Application, K.U., Kurukshetra, INDIA
ckverma@rediffmail.com
ABSTRACT: Multimodal biometric system uses more than one biometric trait for authentication to individual. Unimodal biometric system uses a single biometric trait (fingerprint, face, hand geometry, iris, retina, voice, gait, signature etc.) for authentication (either verification or identification) of user. Unimodal biometric system have some limitations like noise in sensed data, intra-class variation, inter-class variations, distinctiveness, spoof attacks etc. So unimodal biometric system is less secure and less reliable. Some of limitations imposed by unimodal biometric system can be overcome by including multiple traits of individuals. Such systems, known as multimodal biometric system, are more reliable and securable due to presence of multiple traits. This paper represents a brief introduction of multimodal biometric system.
KEYWORDS: Fusion Levels, Multimodal Biometric System, System Modules.
1. INTRODUCTION
Multimodal biometric is a technique by which two modalities (for e.g. face and fingerprint) of a single user combined to acquire higher secure and reliable authentication of the user. For example, before one year
while making of aadhar card, system asked us to give our iris, face and fingerprint as for identification. The combination of these two or more types of biometric information trait, this information keep the advantage of the proficiency of each individual biometric trait and help to meet rigid performance requirements,
reliability, security and robustness against attacks. As shown below figure 1, two biometric traits are used for verification and identification. One of the traits is fingerprint that acquires by a sensor and extracts its feature set for matching it to stored template of fingerprint in the database. Another trait is the face that acquires by a digital camera and extracts its landmarks then match with stored face templates in the database. After matching process finish there should be fusion strategies to combine both decision and put a perfect result to system or individual. Multimodal biometric system tries to provide higher security, reliability, acceptability, easy to measure and less circumvention. Multimodal and Multibiometric systems both are different as Multibiometric system use more than one phase of a single biometric trait (example:left face,
right face, centre face) on the other hand multimodal system uses more than one biometric trait of an individual (example: face and fingerprint).
Fingerprint template
Database
Match-ing
Fusion and decision
Sensor Feature Extractor Accept
Or 
Reject

















Match-ing
Sensor Feature Extractor
Face template
Database
Fig 1: Multimodal Biometric System
To overcome limitations of unimodal; multimodal biometric system provides a higher security and reliability over unimodal biometric system [2].