Journal of forensic sciences最新文献

Of the various rules establishing a mental health clinician's legal duty to take precautions to protect their patient from harming others, the most common is the specificity rule that limits the protective duty to warn reasonably identifiable victims. The specificity rule is important wherein the main or only specified protective measure is warning the victim. In the last quarter century, Pennsylvania adopted the specificity rule from its Supreme Court Emerich decision. In its recent Maas decision, the Pennsylvania Supreme Court expanded the duty to apply to potential victims who are unnamed and unidentifiable except for living on the same floor as the patient of a multiunit building. Victims constituted a group referenced by the patient as a “neighbor,” but from the patient's threats both narrower “next door neighbor” and broader “anyone.” We place this judicial expansion of the duty to warn within the context of professional ethics guidelines and state Tarasoff statutes that pertain to psychiatrists. The potential adverse consequences of this vague expansion of the specificity rule for clinicians, psychiatric patients, and unconnected citizens of Pennsylvania and for other jurisdictions in which courts could misguidedly follow this expansionist example are discussed, along with potential solutions.

The above article, first published online 31^st October 2023, on Wiley Online Library (wileyonlinelibrary.com), has been retracted at the request of the authors, and by agreement between the Editor in Chief, Michael Peat PhD, the American Academy of Forensic Science, and John Wiley & Sons Inc. This retraction has been made because of the omission of the authorship of two co-investigators and premature submission of the manuscript that prevented corrections from, and approval by, the full research team. There was also a departure from the authorizing committee's terms for acknowledgment.

Training is an essential component of onboarding new hires in forensic science service provider (FSSP) laboratories. There are several DNA training standards published by the American Academy of Forensic Sciences (AAFS) Academy Standards Board (ASB) American National Standards Institute (ANSI) accredited framework. In this study, we conducted a survey of forensic DNA laboratory training programs to better understand training activities and materials. The survey was approved by the IRB and emailed to forensic laboratory directors, assistant directors, and/or DNA technical leaders and responses were submitted by them or their designees. Over thirty leaders and stakeholders responded. In this article, we report on the results of the survey. Respondents indicated that training activities included readings, writing assignments and quizzes, shadowing analysts, and mock casework and that training is documented and is a collaborative effort of the technical leader, unit supervisor, advanced forensic scientists, and other analysts and technicians. Laboratories assess competency using multiple methods including performance on mock casework, report writing, laboratory practical and competency tests, and a mock trial and testimony. The top three training activities reported are hands-on practice, shadowing, and readings. The top three focuses of the training are laboratory techniques (extraction, quantitation, amplification, and capillary electrophoresis), STR typing, and mixture analysis. Ethics violations and failure to pass the competency tests and mock trial, even after remediation, are the top reasons new hires fail training. Finally, the top items respondents would like to spend more time offering training on are troubleshooting, validation, and root cause analysis.

Silveira, D, Kendell, A, Shook, B. The impacts of thermocouple insulation failure on the accuracy of temperature measurement data in forensic fire-death scenarios—Part II: Low electrical resistance and contamination. J Forensic Sci 2024; 69(5): 1906–17. https://doi.org/10.1111/1556-4029.15577

In both the Abstract and Highlights for this article, the upper limit for the range of temperature errors that occurred during thermocouple measurements of an ice bath at 0°C was given as “almost 2200°C.” This value was originally measured in Fahrenheit and was not properly converted to Celsius. The correct value in both instances should appear as “almost 1200°C.”

We apologize for this error.

The identification of genital injuries during a medical-forensic examination may impact the investigation and prosecution of sexual assault. The purpose of this meta-analysis was to compare prevalence of genital lacerations (or tears) in persons reporting consensual versus non-consensual vaginal penetration. We hypothesized there would be greater prevalence of genital injuries in the non-consensual group. We searched PubMed, CINAHL, Web of Science, and gray literature for relevant observational case–control studies. Relative risk (RR) ratios using Mantel–Haenszel method were calculated to compare prevalence of genital lacerations between patients reporting consensual versus non-consensual penetration. Mantel–Haenszel chi-square tested significance and Cochran's Q determined between-study heterogeneity. In all studies, genital injuries were observed in both groups. The overall RR of 1.26 indicated a 26% greater risk of genital lacerations in the non-consensual group compared to the consensual group (p = 0.003). Study heterogeneity was 54% (p = 0.05). Findings suggest a greater likelihood of observing genital lacerations in persons reporting non-consensual vaginal penetration. High heterogeneity reflects variations between studies regarding provider education and training, visualization techniques, patient characteristics, time between assault and examination, and documentation practices. A major limitation is the lack of observer blinding creating potential bias. Findings support the need to standardize genital visualization techniques and documentation in sexual assault medical forensic examinations. Healthcare providers conducting medical forensic examinations must understand the limitations of studies comparing non-consensual and consensual injury findings if asked to provide information or testimony on this topic.

Deepfake videos are growing progressively more competent because of the rapid advancements in artificial intelligence and deep learning technology. This has led to substantial problems around propaganda, privacy, and security. This research provides an analytically novel method for detecting deepfake videos using temporal discrepancies of the various statistical features of video at the pixel level, followed by a deep learning algorithm. To detect minute aberrations typical of deepfake manipulations, this study focuses on both spatial information inside individual frames and temporal correlations between subsequent frames. This study primarily provides a novel Euclidean distance variation probability score value for directly commenting on the authenticity of a deepfake video. Next, fine-tuning of InceptionResNetV2 with the addition of a dense layer is trained FaceForensics++ for deepfake detection. The proposed fine-tuned model outperforms the existing techniques as its testing accuracy on unseen data outperforms the existing methods. The propsd method achieved an accuracy of 99.80% for FF++ dataset and 97.60% accuracy for CelebDF dataset.

A forensic investigator performing source identification on a questioned image from a crime aims to identify the unknown camera that acquired the image. On the camera sensor, minute spatial variations in intensities between pixels, called photo response non-uniformity (PRNU), provide a unique and persistent artifact appearing in every image acquired by the digital camera. This camera fingerprint is used to produce a score between the questioned image and an unknown camera using a court-approved camera identification algorithm. The score is compared to a fixed threshold to determine a match or no match. Error rates for the court-approved camera-identification PRNU algorithm were established on a very large set of image data, making no distinction between images with different brightness levels. Camera exposure settings and in-camera processing strive to produce a visually pleasing image, but images that are too dark or too bright are not uncommon. While prior work has shown that exposure settings can impact the accuracy of the court-approved algorithm, these settings are often unreliable in the image metadata. In this work, we apply the court-approved PRNU algorithm to a large data set where images are assigned a brightness level as a proxy for exposure settings using a novel classification method and then analyze error rates. We find statistically significant differences between error rates for nominal images and for images labeled dark or bright. Our result suggests that in court, the error rate of the PRNU algorithm for a questioned image may be more accurately characterized when considering the image brightness.

A critical component of the biological profile involves accurate estimation of the age-at-death of the decedent(s). While bias, inaccuracy, and population specificity have plagued age estimation methods, these issues are minimized by Bayesian statistics. Our primary analysis generated ages of transition from the published Suchey-Brooks pubic symphysis dataset and coupled them with informative priors derived from two modern American forensic samples (Forensic Data Bank [FDB] and Forensic Anthropology Database for Assessing Methods Accuracy [FADAMA]) to test the accuracy of Bayesian analysis against the original method. Exact binomial tests assessed the accuracy of the generated age ranges; realized accuracies and bias are reported for final age ranges at various coverages. In a second analysis, the Bayesian Suchey-Brooks parameters and original, non-Bayesian age ranges were also applied to FADAMA. In the primary analyses, the Bayesian approach improved age estimates over traditional ranges, especially for females. Highest posterior density ranges at 95% provided realized accuracies on a holdout sample between 93% and 96% with extremely low bias for most phases. We provide lookup tables with Bayesian age ranges for various coverages. In the second analysis, realized accuracies were slightly higher in the non-Bayesian approach for both sexes (86%–92% vs. 83%–91%), due to lower precision and likely practitioner bias in aging forensic cases. The popularity of the Suchey-Brooks pubic symphysis aging method in modern forensic casework necessitates the use of the Bayesian approach and we encourage practitioners to utilize the lookup tables for forensic casework in the United States.

In the fingermark detection field, background illustrations may negatively impact the visibility of the marks in presence. They can indeed locally reduce the contrast or mask ridge details. When conducting a research project, the choice for plain or illustrated substrates is mostly driven by the research objectives as well as the development level of the investigated technique (e.g., new reagent, comparison between two techniques, semi-operational test). Also, it is not clear how the human assessors in charge of grading the quality of the detected fingermarks actually cope with fingermarks standing astride background prints. In recent publications, latent quality metrics (LQM) and improved latent fingerprint quality metrics (ILFQM) algorithms have been identified as potential alternatives to traditional human assessment. The authors of the present study were therefore interested in assessing the performance of these algorithms with regards to illustrated backgrounds. To reach that goal, 480 fingermarks were deposited on porous (240) and non-porous (240) items presenting various kinds of illustrations (e.g., colored areas, writings, geometric shapes, lines). Several detection techniques were then applied in sequence: 1,2-indanedione/zinc + ninhydrin + physical developer for porous items, and superglue fuming followed by dye-staining or black magnetic powder for non-porous items. A total of 884 pictures were obtained and their quality assessed by LQM and ILFQM. The results were then analyzed to determine if background prints may influence the metrics provided both algorithms, and if yes, to which extent.