Latent fingerprints are one of the most crucial sources of evidence in forensic investigations. As such, development of automatic latent fingerprint recognition systems to quickly and accurately identify the suspects is one of the most pressing problems facing fingerprint researchers. One of the first steps in manual latent processing is for a fingerprint examiner to perform a triage by assigning one of the following three values to a query latent: Value for Individualization (VID), Value for Exclusion Only (VEO), or No Value (NV). However, latent value determination by examiners is known to be subjective, resulting in large intra-examiner and inter-examiner variations. Furthermore, in spite of the guidelines available, the underlying bases that examiners implicitly use for value determination are unknown. In this paper, we propose a crowdsourcing based framework for understanding the underlying bases of value assignment by fingerprint examiners, and use it to learn a predictor for quantitative latent value assignment. Experimental results are reported using four latent fingerprint databases, two from forensic casework (NIST SD27 and MSP) and two collected in laboratory settings (WVU and IIITD), and a state-of-the-art latent automated fingerprint identification system (AFIS). The main conclusions of this paper are as follows: 1) crowdsourced latent value is more robust than prevailing value determination (VID, VEO, and NV) and latent fingerprint image quality for predicting AFIS performance; 2) two bases can explain expert value assignments, which can be interpreted in terms of latent features; and 3) our value predictor can rank a collection of latents from most informative to least informative.

Existing statistical methods for estimating the log-likelihood ratio from biometric scores include parametric estimation, kernel density estimation, and recently adopted logistic regression estimation. There has been a growing interest to study the repeatability and reproducibility of these methods on biometric datasets after the 2009 National Research Council report [15] and the 2016 President's Council of Advisors on Science and Technology report [1]. For a statistical forensic evaluation method to be repeatable, it needs to generate consistent log-likelihood ratios for various sample size ratios between the genuine (mated) and imposter (non-mated) scores from the same database. It is a well known fact, that for logistic regression methods, the estimated intercept value depends on the sample size ratio between the two groups. Therefore, when computing log-likelihood ratios using logistic regression estimation, different genuine and impostor sample size ratios could result in different log-likelihood ratio values. We performed extensive simulations and used face and fingerprint biometric datasets to investigate repeatability and reproducibility of existing log-likelihood ratio estimation methods.

We report performance of one-to-many fingerprint identification algorithms using one, two, four, eight or ten fingers for recognition. Performance is quantified in terms of recognition accuracy (false positive and false negative identification rate), robustness of algorithms (failure to process images) and computational efficiency (time to execute and size of the generated templates). We measured performance on a large dataset of fingerprint images of operational quality using varying enrollment sizes. The main contributions of this paper are two fold: a) open-set identification performance measures and b) an assessment of core capability of current one-to-many fingerprint recognition algorithms. This is accomplished using a subset of algorithms reported in the Fingerprint Vendor Technology Evaluation (FpVTE2012) that show the range and bounds of performance seen in that evaluation.

Image-specific false match and false non-match error rates are defined by inheriting concepts from the biometric zoo. These metrics support failure mode analyses by allowing association of a covariate (e.g., dilation for iris recognition) with a matching error rate without having to consider the covariate of a comparison image. Image-specific error rates are also useful in detection of ground truth errors in test datasets. Images with higher image-specific error rates are more ``difficult'' to recognize, so these metrics can be used to assess the level of difficulty of test corpora or partition a corpus into sets with varying level of difficulty. Results on use of image-specific error rates for ground-truth error detection, covariate analysis and corpus partitioning is presented.

We document methods for the quantitative evaluation of systems that produce a scalar summary of a biometric sample's quality. We are motivated by a need to test claims that quality measures are predictive of matching performance. We regard a quality measurement algorithm as a black box that converts an input sample to an output scalar. We evaluate it by quantifying the association between those values and observed matching results. We advance detection error trade-off and error versus reject characteristics as metrics for the comparative evaluation of sample quality measurement algorithms. We proceed this with a definition of sample quality, a description of the operational use of quality measures. We emphasize the performance goal by including a procedure for annotating the samples of a reference corpus with quality values derived from empirical recognition scores.

This paper presents an R&D framework used by the National Institute of Standards and Technology (NIST) for biometric technology testing and evaluation. The focus of this paper is on fingerprint-based verification and identification. Since 9/11 the NIST Image Group has been mandated by Congress to run a program for biometric technology assessment and biometric systems certification. Four essential areas of activity are discussed: 1) developing test datasets; 2) conducting performance assessment; 3) technology development; and 4) standards participation. A description of activities and accomplishments are provided for each of these areas. In the process, methods of performance testing are described and results from specific biometric technology evaluations are presented. This framework is anticipated to have broad applicability to other technology and application domains

We present a novel measure of fingerprint image quality, which can be used to estimate fingerprint match performance. This means presenting the matcher with good quality fingerprint images will result in high matcher performance, and vice versa, the matcher will perform poorly for poor quality fingerprints. We discuss the implementation of our fingerprint image quality metric and we present the results of testing it on 280 different combinations of fingerprint image data and fingerprint matcher systems. We found that the metric predicts matcher performance for all systems and datasets. Our definition of quality can be applied to other biometric modalities and upon proper feature extraction can be used to assess quality of any mode of biometric samples.