Quantitative matching of forensic evidence fragments using fracture surface topography and statistical learning.

The complex jagged trajectory of fractured surfaces of two pieces of forensic evidence is used to recognize a "match" by using comparative microscopy and tactile pattern analysis. The material intrinsic properties and microstructures, as well as the exposure history of external forces on a fragment of forensic evidence have the premise of uniqueness at a relevant microscopic length scale (about 2-3 grains for cleavage fracture), wherein the statistics of the fracture surface become non-self-affine. We utilize these unique features to quantitatively describe the microscopic aspects of fracture surfaces for forensic comparisons, employing spectral analysis of the topography mapped by three-dimensional microscopy. Multivariate statistical learning tools are used to classify articles and result in near-perfect identification of a "match" and "non-match" among candidate forensic specimens. The framework has the potential for forensic application across a broad range of fractured materials and toolmarks, of diverse texture and mechanical properties.

Tracking the growth of visual evidence in fingerprint comparison tasks.

Fingerprint comparisons are extended in time due to the fine details (minutiae) that necessitate multiple eye fixations throughout the comparison. How is evidence accumulated across these multiple regions? The present work measures decisions at multiple points during a comparison to address how feature diagnosticity and image clarity play a role in evidence accumulation. We find that evidence is accumulated at a constant rate over time, with evidence for identification and exclusion accumulated at similar rates. Manipulations of image diagnosticity and image clarity demonstrate two exceptions to this constant rate: Highly diagnostic evidence followed by weak evidence tends to lose the initial benefits of the strong start, and low image clarity at the start of the comparison can be overcome with high image clarity at the end of the comparison. The results suggest that examiners tend to treat each region fairly independently (as demonstrated by linear evidence accumulation), with only weak evidence for hysteresis effects that tend to fade as additional regions are presented. Data from transition probability matrices support an incremental evidence accumulation account, with very little evidence for rapid "aha" moments even for exclusion decisions. The results are consistent with a model in which each fixated region contributes an independent unit of evidence, and these accumulate to form an eventual decision. Fingerprint comparisons do not seem to depend on which regions are selected first, and thus examiners need not worry about finding the most diagnostic region first, but instead focus on conducting a complete analysis of the latent print.

The benefits of errors during training.

Errors are generally not thought of as a positive thing - not in society at large, and especially not in forensic science. However, there is a large body of literature in the field of cognitive science (particularly from psychology and education research) that highlights the benefits that can be gained from using errors made in training to improve learning. Although none of these studies was done directly in the forensic science disciplines, there are nonetheless lessons to be learned about how errors may most effectively be used to maximize their benefits to learning. This article presents an overview of the literature on learning from errors and suggests principles that may be of benefit to forensic science today, as well as suggesting areas where specific research may be of benefit to forensic science in the future.

Validating strength-of-support conclusion scales for fingerprint, footwear, and toolmark impressions.

In the pattern comparison disciplines such as fingerprints, footwear, and toolmarks, the results of a comparison are communicated by examiners in the form of categorical conclusions such as Identification or Exclusion. These statements have been criticized as requiring knowledge of prior probabilities by the examiners and being overinterpreted by laypersons. Alternative statements based on strength-of-support language have been proposed. The current study compares traditional conclusion scales against strength-of-support scales to determine how these new statements might be used by examiners in casework. Each participant completed 60 comparisons within their discipline, which were designed to approximate casework conditions, using either a traditional or a strength-of-support conclusion scale. The scale used on each trial was randomly assigned, and participants knew the scale for that trial as they began the comparison. Fingerprint examiners were much less likely to use Extremely Strong Support for Common Source than Identification. Footwear examiners treated the traditional and strength-of-support scales similarly, but toolmark examiners were much less likely to use Extremely Strong Support for Common Source than Identification, similar to fingerprint examiners. A separate group of fingerprint examiners used Identification less often when an expanded scale was available. The results demonstrate that expanded scales may result in the highest conclusion category being used less often by examiners when other alternatives are possible, and the term "extremely strong support" may introduce risk aversion on the part of examiners.

Quantitative matching of forensic evidence fragments utilizing 3D microscopy analysis of fracture surface replicas.

Silicone casts are widely used by practitioners in the comparative analysis of forensic items. Fractured surfaces carry unique details that can provide accurate quantitative comparisons of forensic fragments. In this study, a statistical analysis comparison protocol was applied to a set of 3D topological images of fractured surface pairs and their replicas to provide confidence in the quantitative statistical comparison between fractured items and their silicone cast replicas. A set of 10 fractured stainless steel samples were fractured from the same metal rod under controlled conditions and were replicated using a standard forensic casting technique. Six 3D topological maps with 50% overlap were acquired for each fractured pair. Spectral analyses were utilized to identify the correlation between topological surface features at different length scales of the surface topology. We selected two frequency bands over the critical wavelength (greater than two-grain diameters) for statistical comparison. Our statistical model utilized a matrix-variate t-distribution that accounts for overlap between images to model match and non-match population densities. A decision rule identified the probability of matched and unmatched pairs of surfaces. The proposed methodology correctly classified the fractured steel surfaces and their replicas with a posterior probability of match exceeding 99.96%. Moreover, the replication technique shows potential in accurately replicating fracture surface topological details with a wavelength greater than 20 µm, which far exceeds the feature comparison range on most metallic alloy surfaces. Our framework establishes the basis and limits for forensic comparison of fractured articles and their replicas while providing a reliable fracture mechanics-based quantitative statistical forensic comparison.

The Utility of Expanded Conclusion Scales During Latent Print Examinations.

During fingerprint comparisons, a latent print examiner visually compares two impressions to determine whether or not they originated from the same source. They consider the amount of perceived detail in agreement or disagreement and accumulate evidence toward same source and different sources propositions. This evidence is then mapped to one of three conclusions: Identification, Inconclusive, or Exclusion. A limitation of this 3-conclusion scale is it can lose information when translating the conclusion from the internal strength-of-evidence value to one of only three possible conclusions. An alternative scale with two additional values, support for different sources and support for common sources, has been proposed by the Friction Ridge Subcommittee of OSAC. The expanded scale could lead to more investigative leads but could produce complex trade-offs in both correct and erroneous identifications. The aim of the present study was to determine the consequences of a shift to expanded conclusion scales in latent print comparisons. Latent print examiners each completed 60 comparisons using one of the two scales, and the resulting data were modeled using signal detection theory to measure whether the expanded scale changed the threshold for an "Identification" conclusion. When using the expanded scale, examiners became more risk-averse when making "Identification" decisions and tended to transition both the weaker Identification and stronger Inconclusive responses to the "Support for Common Source" statement. The results demonstrate the utility of an expanded conclusion scale and also provide guidance for the adoption of these or similar scales.

Characterizing Human Expertise Using Computational Metrics of Feature Diagnosticity in a Pattern Matching Task.

Forensic evidence often involves an evaluation of whether two impressions were made by the same source, such as whether a fingerprint from a crime scene has detail in agreement with an impression taken from a suspect. Human experts currently outperform computer-based comparison systems, but the strength of the evidence exemplified by the observed detail in agreement must be evaluated against the possibility that some other individual may have created the crime scene impression. Therefore, the strongest evidence comes from features in agreement that are also not shared with other impressions from other individuals. We characterize the nature of human expertise by applying two extant metrics to the images used in a fingerprint recognition task and use eye gaze data from experts to both tune and validate the models. The Attention via Information Maximization (AIM) model (Bruce & Tsotsos, 2009) quantifies the rarity of regions in the fingerprints to determine diagnosticity for purposes of excluding alternative sources. The CoVar model (Karklin & Lewicki, 2009) captures relationships between low-level features, mimicking properties of the early visual system. Both models produced classification and generalization performance in the 75%-80% range when classifying where experts tend to look. A validation study using regions identified by the AIM model as diagnostic demonstrates that human experts perform better when given regions of high diagnosticity. The computational nature of the metrics may help guard against wrongful convictions, as well as provide a quantitative measure of the strength of evidence in casework.

The impact of fatigue on latent print examinations as revealed by behavioral and eye gaze testing.

Eye tracking and behavioral methods were used to assess the effects of fatigue on performance in latent print examiners. Eye gaze was measured both before and after a fatiguing exercise involving fine-grained examination decisions. The eye tracking tasks used similar images, often laterally reversed versions of previously viewed prints, which holds image detail constant while minimizing prior recognition. These methods, as well as a within-subject design with fine grained analyses of the eye gaze data, allow fairly strong conclusions despite a relatively small subject population. Consistent with the effects of fatigue on practitioners in other fields such as radiology, behavioral performance declined with fatigue, and the eye gaze statistics suggested a smaller working memory capacity. Participants also terminated the search/examination process sooner when fatigued. However, fatigue did not produce changes in inter-examiner consistency as measured by the Earth Mover Metric. Implications for practice are discussed.

Temporal sequences quantify the contributions of individual fixations in complex perceptual matching tasks.

Perceptual tasks such as object matching, mammogram interpretation, mental rotation, and satellite imagery change detection often require the assignment of correspondences to fuse information across views. We apply techniques developed for machine translation to the gaze data recorded from a complex perceptual matching task modeled after fingerprint examinations. The gaze data provide temporal sequences that the machine translation algorithm uses to estimate the subjects' assumptions of corresponding regions. Our results show that experts and novices have similar surface behavior, such as the number of fixations made or the duration of fixations. However, the approach applied to data from experts is able to identify more corresponding areas between two prints. The fixations that are associated with clusters that map with high probability to corresponding locations on the other print are likely to have greater utility in a visual matching task. These techniques address a fundamental problem in eye tracking research with perceptual matching tasks: Given that the eyes always point somewhere, which fixations are the most informative and therefore are likely to be relevant for the comparison task?

Behavioral and electrophysiological evidence for configural processing in fingerprint experts.

Visual expertise in fingerprint examiners was addressed in one behavioral and one electrophysiological experiment. In an X-AB matching task with fingerprint fragments, experts demonstrated better overall performance, immunity to longer delays, and evidence of configural processing when fragments were presented in noise. Novices were affected by longer delays and showed no evidence of configural processing. In Experiment 2, upright and inverted faces and fingerprints were shown to experts and novices. The N170 EEG component was reliably delayed over the right parietal/temporal regions when faces were inverted, replicating an effect that in the literature has been interpreted as a signature of configural processing. The inverted fingerprints showed a similar delay of the N170 over the right parietal/temporal region, but only in experts, providing converging evidence for configural processing when experts view fingerprints. Together the results of both experiments point to the role configural processing in the development of visual expertise, possibly supported by idiosyncratic relational information among fingerprint features.