An introduction to the good, the bad, & the ugly face recognition challenge problem

Abstract

The Good, the Bad, & the Ugly Face Challenge Problem was created to encourage the development of algorithms that are robust to recognition across changes that occur in still frontal faces. The Good, the Bad, & the Ugly consists of three partitions. The Good partition contains pairs of images that are considered easy to recognize. On the Good partition, the base verification rate (VR) is 0.98 at a false accept rate (FAR) of 0.001. The Bad partition contains pairs of images of average difficulty to recognize. For the Bad partition, the VR is 0.80 at a FAR of 0.001. The Ugly partition contains pairs of images considered difficult to recognize, with a VR of 0.15 at a FAR of 0.001. The base performance is from fusing the output of three of the top performers in the FRVT 2006. The design of the Good, the Bad, & the Ugly controls for pose variation, subject aging, and subject "recognizability." Subject recognizability is controlled by having the same number of images of each subject in every partition. This implies that the differences in performance among the partitions are result of how a face is presented in each image.

Open source biometric recognition

“…The CSU baseline algorithm suite includes two recently published algorithms, local region PCA (LRPCA) [17] and LDA with color spaces and cohort normalization (CohortLDA) [16]. The framework is written in Python and R with installation instructions provided for Mac and Windows.…”

“…The framework is written in Python and R with installation instructions provided for Mac and Windows. Scripts are included to run the algorithms against the GBU [17] and LFW [8] datasets, though programming language requirements and the lack of a well defined API make it difficult to incorporate their work into new applications. Source code is released as a zipped archive and it is unclear how developers should contribute back to the project.…”

“…Three criteria are applied that gauge if a project (i) implements modern algorithms, (ii) is under active development, and (iii) is readily deployable in applications. Two particularly notable solutions from Colorado State University (CSU) [17] and OpenCV [4] are discussed here further.…”

Measuring and mitigating targeted biometric impersonation

“…This section evaluates the effects of targeted attacks on the CSU baseline algorithm developed by Bolme et al [15] for the good, the bad and the ugly face recognition challenge [16]. The system has been trained using images from the NIST multiple biometric grand challenge data set [17].…”

An Open Source Framework for Standardized Comparisons of Face Recognition Algorithms

“…To show their advantage over other existing algorithms, face recognition experiments have typically been executed on one or more publicly available facial image databases [1,2,3,4,5,6,7,8]. Unfortunately, often these databases are not accompanied by strict experimental protocols or the protocols that are provided are biased.…”