Quality Estimation for Facial Biometrics

Quality Estimation for Facial Biometrics

Sanjida Nasreen Tumpa (University of Calgary, Canada), Andrei Dmitri Gavrilov (University of British Columbia, Canada), Omar Zatarain Duran (University of Calgary, Canada), Fatema Tuz Zohra (University of Calgary, Canada) and Marina L. Gavrilova (University of Calgary, Canada)
DOI: 10.4018/978-1-7998-3038-2.ch014
OnDemand PDF Download:
No Current Special Offers


Over past decade, behavioral biometric systems based on face recognition became leading commercial systems that meet the need for fast and efficient confirmation of a person's identity. Facial recognition works on biometric samples, like image or video frames, to recognize people. The performance of an automated face recognition system has a strong relationship with the quality of the biometric samples. In this chapter, the authors propose a quality estimation method based on a linear regression analysis to characterize the relationship between different quality factors and the performance of a face recognition system. The regression model can predict the overall quality of a facial sample which reflects the effects of various quality factors on that sample. The authors evaluated the quality estimation model on the Extended Yale Database B, finally formulating a data set of samples which will enable efficient implementation of biometric facial recognition.
Chapter Preview


Biometrics is the science of recognizing a person based on the unique physiological or behavioral characteristics. Facial biometrics is one of the popular branches of biometrics which refers to an authentication technology that can recognize and verify human based on facial characteristics. One of the key advantages of facial biometrics over other biometrics is, it does not require the cooperation of the subject to identify. Sometimes the subject may remain fully unaware of the presence of the identification system. Hence, the significance of facial biometrics to ensure public security is high. This technology is mostly used in airports, mass gatherings, multiplexes, public places, and other crowded areas. Generally, the identification and verification are done based on the digital images or video frame obtained from video sources. Therefore, the quality of the digital images is closely connected with the performance of a face recognition system (Grother&Tabassi 2007) (Alonso-Fernandez et al. 2012) (Bharadwaj et al. 2014) (Khryashchev et al. 2018) (Yu et al. 2018) (Hernandez-Ortega et al. 2019) (Vignesh et al. 2015). However, in the real-time video surveillance scenarios, the face images may suffer from high distortion due to uncooperative users and uncontrolled environments. For example, the captured image can be of low resolution, blurry, noisy, etc. which may create a great challenge for the face recognition system. The performance of a face recognition system degrades with poor quality of data due to variations in lighting conditions, facial expressions, pose, resolution, camera movement, occlusion, and poor sensor quality (Sellahewa&Jassim 2010) (Abaza et al. 2012) (Abaza et al. 2014) (Punnappurath et al. 2015) (Khryashchev et al. 2018) (Vignesh et al. 2015).

The goal of the Biometric Quality Assessment (BQA) technique is to determine the capability of an image to qualify as a biometric sample. There have been previous attempts to structure face-based quality factors based on digital formatting of facial images, scenes, and photographs. Several techniques have been previously suggested to estimate data quality (Wang & Bovik 2002) (Abaza et al. 2012) (Abaza et al. 2014) (Sang et al. 2009) (Bosse et al. 2017) (Olsen et al. 2016). However, only very few researchers considered various quality factors jointly (Abaza et al. 2012) (Abaza et al. 2014). Thus, this book chapter describes effective quality estimation model, which can characterize the quality of the biometric image by integrating different quality measures into a single quality score which is an indicator of the overall quality of the facial image. The model also considers the effect of various quality factors on face recognition performance.

Complete Chapter List

Search this Book: