SNoW: Understanding the causes of strong, neutral, and weak face impostor pairs

Sgroi, Amanda; Bowyer, Kevin W.; Flynn, Patrick J.; Phillips, P. Jonathon

doi:10.1109/btas.2013.6712697

Cited by 4 publications

(7 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This constraint requires that when selecting the partitions, images in a pair must have been acquired on different days. In [3], the different days constraint was utilized for creating the SNoW partitions.…”

Section: The Snow Partitioning Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

Strong, Neutral, or Weak: Exploring the Impostor Score Distribution

Sgroi

Flynn

Bowyer

et al. 2015

IEEE Trans.Inform.Forensic Secur.

Self Cite

View full text Add to dashboard Cite

The strong, neutral, or weak (SNoW) face impostor pairs problem is intended to explore the causes and impact of impostor face pairs that are inherently strong (easily recognized as nonmatches) or weak (possible false matches). The SNoW technique develops three partitions within the impostor score distribution of a given data set. Results provide evidence that varying degrees of impostor scores impact the overall performance of a face recognition system. This paper extends our earlier work to incorporate improvements regarding outlier detection for partitioning, explores the SNoW concept for the additional modalities of fingerprint and iris, and presents methods for how to begin to reveal the causes of weak impostor pairs. We also show a clear operational difference between strong and weak comparisons as well as identify partition stability across multiple algorithms.

show abstract

Section: The Snow Partitioning Algorithmmentioning

confidence: 99%

“…In this paper, we expand our initial work on studying the tail of the impostor distribution [3]. Analysis of these scores and the impostor distribution as a whole has the potential to help us understand how to characterize image pairs that have the most authenticlike nonmatch scores.…”

Section: Introductionmentioning

confidence: 98%

Strong, Neutral, or Weak: Exploring the Impostor Score Distribution

Sgroi

Flynn

Bowyer

et al. 2015

IEEE Trans.Inform.Forensic Secur.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Through the use of multiple algorithms, we consider pairs which tend to be more match-like or more nonmatch-like across multiple comparison methods. To determine which scores are more or less matchlike, we use the SNoW partitioning framework to identify weak and strong nonmatch pairs for all methods and then identify which pairs are always partitioned as weak or strong given the four comparison methods [19]. [8].…”

Section: Preliminariesmentioning

confidence: 99%

“…The SNoW partitioning algorithm partitions the impostor distribution into three components [19]. The Strong partition contains image pairs that are easy to categorize as nonmatches.…”

Section: Snow Partitioningmentioning

confidence: 99%

“…We use the SNoW partitioning technique with a combination of three dataset and four matching algorithms to discern between strong impostor scores (least match-like) and weak impostor scores (more match-like) [19]. Given these score sets, we then analyze the score distributions for strong and weak scores which have zero, one, two, three, or four matching metadata values.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Metadata-based understanding of impostor pair score variations

Sgroi

Bowyer

Flynn

2014

2014 IEEE International Workshop on Information Forensics and Security (WIFS)

Self Cite

View full text Add to dashboard Cite

Metadata about a given face image can include information such as the subject's year of birth, subject gender, and date of acquisition. By determining the degree of metadata matches between the gallery and probe images (such as two subjects having the same gender) we hypothesize that more metadata values that match for an impostor image pair increases the likelihood of a false match. In this work, we explore year of birth, gender, date of acquisition, and expression in an attempt to understand variations in match scores produced by impostor image pairs. Impostor pairs that fall in the weak partition as identified by the previously developed SNoW technique have a slightly larger number of matching metadata values. However, there is little to no statistically significant difference in the scores produced by image pairs with more matching metadata between strong and weak impostor pairs.

show abstract