scholarly journals Discriminative power of features used by forensic document examiners in the analysis of handwriting

2006 ◽  
Author(s):  
Vladimir Pervouchine
Keyword(s):  
Discriminative Power ◽  
Document Examiners

Seeking Patterns in the Forensic Analysis of Handwriting and Speech

2008 ◽  
pp. 110-139
Author(s):  
Graham Leedham ◽  
Vladimir Pervouchine ◽  
Haishan Zhong
Keyword(s):  
Forensic Analysis ◽  
High Accuracy ◽  
Discriminative Power ◽  
Person Identification ◽  
The Future ◽  
Quantitative Manner ◽  
Speech Imitation ◽  
Document Examiners

This chapter examines features of handwriting and speech and their effectiveness at determining whether the identity of a writer or speaker can be identified from his or her handwriting or speech. For handwriting, some of the subjective and qualitative features used by document examiners are investigated in a scientific and quantitative manner based on the analysis of three characters (“d,” “y,” and “f”) and the grapheme “th.” For speech, several frequently used features are compared for their strengths and weaknesses in distinguishing speakers. The results show that some features do have good discriminative power, while others are less effective. Acceptable performance can be obtained in many situations using these features. However, the effect of handwriting forgery/disguise or conscious speech imitation/alteration on these features is not investigated. New and more powerful features are needed in the future if high accuracy person identification can be achieved in the presence of disguise or forgery.

Investigating the Potential for Training Context Effects to Influence Forensic Document Examiners' Relative Skill at Writer Individualization and Exclusion

2015 ◽  
Vol 25 ◽  
pp. 27-35 ◽  
Author(s):  
Tonya Trubshoe ◽  
Bryan Found
Keyword(s):  
Context Effects ◽  
Test Results ◽  
Training Context ◽  
Training Set ◽  
Blind Trial ◽  
Relative Ability ◽  
Trial Test ◽  
Case Files ◽  
Document Examiners

The relative ability of forensic document examiners (FDEs) to provide support for the proposition of individualization or exclusion on the basis of handwriting features was investigated by surveying opinions expressed in case files by one laboratory’s FDEs and comparing this data to blind trial test results taken over a five year period. The survey of FDEs opinions on reports showed that opinions were skewed towards support for writer individualization over writer exclusion 92% of the time. Since historically FDEs develop their skills with respect to individualization/exclusion primarily on case files, it is proposed that this unbalanced training context may skew their abilities to carry out the tasks. To determine one laboratory’s capacity to correctly provide both individualization and exclusion evidence, results of blind validation trials were analyzed. For natural writing written and not written by the specimen writer, FDEs were 62 times more inconclusive when providing support for exclusion of the specimen writer when the specimen writer did not author the questioned sample, than they were for providing support for individualization when the specimen writer wrote the questioned sample. An intriguing possibility is that because of the unbalanced training set, government FDEs may acquire skills which are skewed towards individualization over exclusion.   Purchase Article - $10

The Frequency of Occurrence of Handwriting Performance Features Used to Predict Whether Questioned Signatures are Simulated

2018 ◽  
Vol 28 ◽  
pp. 35-42
Author(s):  
David Black ◽  
Bryan Found ◽  
Doug Rogers
Keyword(s):  
Large Population ◽  
Quality Parameters ◽  
Frequency Of Occurrence ◽  
Relative Speed ◽  
Simulation Process ◽  
Spatial Features ◽  
And Performance ◽  
Document Examiners ◽  
Line Trace

Forensic Document Examiners (FDEs) examine the physical morphology and performance attributes of a line trace when comparing questioned to specimen handwriting samples for the purpose of determining authorship. Along with spatial features, the elements of execution of the handwriting are thought to provide information as to whether or not a questioned sample is the product of a disguise or simulation process. Line features such as tremor, pen-lifts, blunt beginning and terminating strokes, indicators of relative speed, splicing and touch ups, are subjectively assessed and used in comparisons by FDEs and can contribute to the formation of an opinion as to the validity of a questioned sample of handwriting or signatures. In spite of the routine use of features such as these, there is little information available regarding the relative frequency of occurrence of these features in populations of disguised and simulated samples when compared to a large population of a single individual’s signature. This study describes a survey of the occurrence of these features in 46 disguised signatures, 620 simulated signatures (produced by 31 different amateur forgers) and 177 genuine signatures. It was found that the presence of splices and touch-ups were particularly good predictors of the simulation process and that all line quality parameters were potentially useful contributors in the determination of the authenticity of questioned signatures. Purchase Article - $10

Effect of Visual Feedback on the Static and Kinematic Characteristics of Handwriting

2017 ◽  
Vol 27 ◽  
pp. 5-21 ◽  
Author(s):  
Michael Pertsinakis
Keyword(s):  
Visual Feedback ◽  
Educational Level ◽  
Movement Analysis ◽  
Individual Characteristics ◽  
Software Module ◽  
Medium Education ◽  
Education Levels ◽  
Pen Tablet ◽  
The Individual ◽  
Document Examiners

Research on visual feedback has not produced consistent results to show how visual feedback or the lack, thereof, influences individual handwriting characteristics. A two-pronged approach was designed to investigate the degree of this influence. For this purpose, samples of signatures as well as cursive and block text, written with and without visual feedback, were collected from 40 volunteers and imported into a PC via a pen tablet, using an electronic inking pen. The data was analyzed in a handwriting movement analysis software module specially designed for this research that was added to the software MovAlyzeR by Neuroscript LLC. Two forensic document examiners (FDEs) independently analyzed samples from the two groups (samples executed with normal visual feedback versus the group of samples executed without visual feedback). They found no fundamental differences between these two groups. Their analyses also demonstrated that a large number of similarities existed in the general design of the allographs (alternative forms of a letter or other grapheme) and in the pictorial aspects, regardless of the complexity of the samples. In the cursive and block handwriting, four main qualitative characteristics were linked to the absence of visual feedback: change of overall size, non-uniformity of left margins, change of baseline alignment, and inclusion of extra trajectories. The statistical analysis verified the above findings. The comparative analysis also suggests that gender, educational level (above high school) and handedness create an insignificant influence on the individual characteristics of writing produced with and without visual feedback. The only notable exception is the relationship between signature duration and educational level. The volunteers with a medium education level showed a significant increase in duration while signing their names without visual feedback in comparison to those with higher education levels. The combination of the above findings suggests that handwriting is not fundamentally influenced by visual feedback.  Purchase Article - $10

scholarly journals Searching for universal model of amyloid signaling motifs using probabilistic context-free grammars

2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Witold Dyrka ◽  
Marlena Gąsior-Głogowska ◽  
Monika Szefczyk ◽  
Natalia Szulc
Keyword(s):  
Functional Relationship ◽  
High Sensitivity ◽  
Alternative Methods ◽  
Discriminative Power ◽  
Context Free Grammar ◽  
Protein Motifs ◽  
Functional Features ◽  
Universal Grammars ◽  
Context Free ◽  
Probabilistic Context

Abstract Background Amyloid signaling motifs are a class of protein motifs which share basic structural and functional features despite the lack of clear sequence homology. They are hard to detect in large sequence databases either with the alignment-based profile methods (due to short length and diversity) or with generic amyloid- and prion-finding tools (due to insufficient discriminative power). We propose to address the challenge with a machine learning grammatical model capable of generalizing over diverse collections of unaligned yet related motifs. Results First, we introduce and test improvements to our probabilistic context-free grammar framework for protein sequences that allow for inferring more sophisticated models achieving high sensitivity at low false positive rates. Then, we infer universal grammars for a collection of recently identified bacterial amyloid signaling motifs and demonstrate that the method is capable of generalizing by successfully searching for related motifs in fungi. The results are compared to available alternative methods. Finally, we conduct spectroscopy and staining analyses of selected peptides to verify their structural and functional relationship. Conclusions While the profile HMMs remain the method of choice for modeling homologous sets of sequences, PCFGs seem more suitable for building meta-family descriptors and extrapolating beyond the seed sample.

The role of cardiopulmonary exercise testing in a contemporary heart failure population

2021 ◽  
Vol 28 (Supplement_1) ◽  
Author(s):  
P Garcia Bras ◽  
A Valentim Goncalves ◽  
J Reis ◽  
T Pereira Da Silva ◽  
R Ilhao Moreira ◽  
...  
Keyword(s):  
Heart Failure ◽  
Risk Stratification ◽  
Exercise Testing ◽  
Predictive Power ◽  
Cardiopulmonary Exercise Testing ◽  
Composite Endpoint ◽  
Left Ventricular ◽  
Left Ventricular Ejection ◽  
Discriminative Power ◽  
Cardiopulmonary Exercise

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Cardiopulmonary exercise testing (CPET) is used for risk stratification in patients with chronic heart failure (CHF). However, there is a lack of information regarding CPET prognostic power in patients under new HF therapies such as sacubitril/valsartan, Mitraclip, IV iron or SGLT2 inhibitors. The aim of this study was to evaluate the prognostic value of CPET parameters in a contemporary subset of patients with optimal medical and device therapy for CHF. Methods Retrospective evaluation of patients with CHF submitted to CPET in a tertiary center. Patients were followed up for 24 months for the composite endpoint of cardiac death, urgent heart transplantation or left ventricular assist device. CPET parameters, including peak oxygen consumption (pVO2) and VE/VCO2 slope, were analysed and their predictive power was measured. HF events were stratified according to cut-off values defined by the International Society for Heart and Lung Transplantation (ISHLT) guidelines: pVO2 of ≤12 mL/Kg/min and VE/VCO2 slope of >35. Results CPET was performed in 204 patients, from 2014 to 2018. Mean age was 59 ± 13 years, 83% male, with a mean left ventricular ejection fraction of 33 ± 8%, and a mean Heart Failure Survival Score of 8.6 ± 1.3. The discriminative power of CPET parameters is displayed in the Table. In patients with pVO2 ≤12 mL/Kg/min, the composite endpoint occurred in 18% of patients. A pVO2 value of ≤12 mL/Kg/min had a positive predictive power of 18% while pVO2 >12 had a negative predictive power of 93%. Regarding VE/VCO2 slope >35, the composite endpoint occurred in 13% of patients. A VE/VCO2 slope value of >35 had a positive predictive power of 13% while VE/VCO2 slope <35 had a negative predictive power or 94%. Conclusion Using ISHLT guideline cut-off values for advanced HF therapies patient selection, there was a reduced number of HF events (<20%) at 24 months in patients under optimal CHF therapy. While pVO2 and VE/VCO2 slope are still valuable parameters in risk stratification, redefining cut-off values may be necessary in a modern HF population. Discriminative power of CPET parameters Parameters HR; 95% CI AUC p-value Peak VO2 0.824 (0.728-0.934) 0.781 0.001 Percent of predicted pVO2 0.942 (0.907-0.978) 0.774 0.002 VE/VCO2 slope 1.068 (1.031-1.106) 0.756 0.008 Cardiorespiratory optimal point 1.118 (1.053-1.188) 0.746 0.004 PETCO2 maximum exercise 0.854 (0.768-0.950) 0.775 0.003 Ventilatory Power 0.358 (0.176-0.728) 0.796 0.002 HR Hazard ratio, AUC: Area under the curve, PETCO2: end-tidal CO2 pressure

scholarly journals Comparison of Ligation-Mediated PCR Methods in Differentiation ofMycobacterium tuberculosisStrains

2014 ◽  
Vol 2014 ◽  
pp. 1-4 ◽  
Author(s):  
Anna Zaczek ◽  
Anna Brzostek ◽  
Arkadiusz Wojtasik ◽  
Anna Sajduda ◽  
Jaroslaw Dziadek
Keyword(s):  
Mycobacterium Tuberculosis ◽  
Small Community ◽  
Discriminative Power ◽  
Epidemiological Analysis ◽  
Reference Strains

Fast and inexpensive identification of epidemiological links between limited number ofMycobacterium tuberculosisstrains is required to initially evaluate hospital outbreaks, laboratory crosscontaminations, and family or small community transmissions. The ligation-mediated PCR methods (LM-PCR) appear sufficiently discriminative and reproducible to be considered as a good candidate for such initial, epidemiological analysis. Here, we compared the discriminative power of the recently developed in our laboratory fast ligation amplification polymorphism (FLAP) method with fast ligation-mediated PCR (FLiP). Verification of the results was based on analyzing a set of reference strains and RFLP-IS6110typing. The HGDI value was very similar for both LM-PCR methods and RFLP-IS6110typing. However, only 52% of strains were correspondingly grouped by both FLiP and FLAP methods. Differentiation by FLAP method demonstrated a limited similarity to IS6110-RFLP (37,7%). As much as 78,7% of strains were grouped identically when differentiated by FLiP and IS6110-RFLP methods. The analysis differentiated 31, 35, and 36 groups when using FLAP, FLiP, and RFLP-IS6110methods, respectively.

Sign in / Sign up

Export Citation Format

Share Document