Evaluation on a keystroke authentication system by keying force incorporated with temporal characteristics of keystroke dynamics

This paper presents the study to develop and evaluate techniques to authenticate valid users, using the keystroke dynamics of a user's PIN number entry on a numerical keypad, with force sensing resistors. Added with two conventional parameter lists of elements, i.e. digraph latency times and key hold times, keying force was chosen as a third element. Two experiments were conducted. The first experiment was to evaluate whether the three types of elements derived from keystrokes have a significant effect for subjects and to examine how trials and session effects generated the variation of the three elements. The second experiment was to demonstrate the system performance by calculating the False Rejection Rate (FRR) and the False Acceptance Rate (FAR) of the system. In the second experiment, a total of 20 keystrokes were recorded from each subject one week after the memorizing session, in order to evaluate the FRR of the system. To evaluate the FAR of the system, the subjects pretended to be impostors, and therefore they repeatedly watched videotaped pass trials made by a valid user as many times as they desired, and tried to imitate the keystroke dynamics of the valid users. The subject's keystrokes were then evaluated on whether they could fool the system. The first experiment, ANOVA revealed that a significant effect of subject was found on each of all three elements. Trial was not significantly affected to digraph latency times and peak force; however, it was significantly affected to key hold times. There was a trend that keystroke dynamics characterized by each element showed reformation of their patterns and reached a steady state over the 10 weeks of experimental sessions. The results of the second experiment showed the average equal error rate to be 2.4%. The results of system performance were compared with those of other studies and concluded that it was difficult to obtain enough information to behave as a perfect impostor by monitoring the videotaped keystrokes.     

AGE AND CHOICE BETWEEN RESPONSES IN A SELF-PACED REPETITIVE TASK

Abstract

When young subjects select between two responses of similar amplitudes, and in the same direction, they can overlap identification of one of two signals with the initial movement of a reach appropriate to either. They thus respond more quickly when selecting between responses of similar amplitudes than when selecting between responses of very different amplitudes. Older subjects are loss able to overlap movement and choice time and so benefit less from a choice between similar as against dissimilar responses. Young subjects respond faster when cycles of responses during the task are alternated than when they are repeated. This tendency appears to be related to guessing strategies earlier described as the ‘ negative recency effect ’ or ‘ gambler 's fallacy ’. The latencies of young subjects' responses appear to be directly dependent on the latencies of immediately preceding responses rather than on the response rate for the sequence as a whole. Latencies of old subjects' responses are affected by both factors. Implications for the design of consoles are discussed.     

Evaluation on a keystroke authentication system by keying force incorporated with temporal characteristics of keystroke dynamics

This paper presents the study to develop and evaluate techniques to authenticate valid users, using the keystroke dynamics of a user's PIN number entry on a numerical keypad, with force sensing resistors. Added with two conventional parameter lists of elements, i.e. digraph latency times and key hold times, keying force was chosen as a third element. Two experiments were conducted. The first experiment was to evaluate whether the three types of elements derived from keystrokes have a significant effect for subjects and to examine how trials and session effects generated the variation of the three elements. The second experiment was to demonstrate the system performance by calculating the False Rejection Rate (FRR) and the False Acceptance Rate (FAR) of the system. In the second experiment, a total of 20 keystrokes were recorded from each subject one week after the memorizing session, in order to evaluate the FRR of the system. To evaluate the FAR of the system, the subjects pretended to be impostors, and therefore they repeatedly watched videotaped pass trials made by a valid user as many times as they desired, and tried to imitate the keystroke dynamics of the valid users. The subject's keystrokes were then evaluated on whether they could fool the system. The first experiment, ANOVA revealed that a significant effect of subject was found on each of all three elements. Trial was not significantly affected to digraph latency times and peak force; however, it was significantly affected to key hold times. There was a trend that keystroke dynamics characterized by each element showed reformation of their patterns and reached a steady state over the 10 weeks of experimental sessions. The results of the second experiment showed the average equal error rate to be 2.4%. The results of system performance were compared with those of other studies and concluded that it was difficult to obtain enough information to behave as a perfect impostor by monitoring the videotaped keystrokes.     

A model of metacognition for bushfire fighters

Large-scale bushfires are complex or macrocognitive decision environments (Klein et al.). They involve many people, such as incident management teams, firefighting crews, and resident communities. Those people can also be geographically dispersed. This means that they need to coordinate their bushfire response efforts and manage multiple, often competing, cognitive demands. To do this, bushfire responders need to use metacognitive skills to regulate their thinking, particularly under stressful high cognitive load conditions. We explored these types of issues with career and volunteer bushfire fighters (Frye and Wearing in J Cogn Technol 16(2): 33–44, 2011). We found that rule-based procedures can sometimes reduce errors, and increase safety, because they reduce cognitive load (e.g., ‘you just do it’). However, fixation on other rules and procedures can increase errors, and erode safety, because people fail to adapt to the current situation (e.g., ‘you really need to think about that’). In this paper we use a model of metacognition to describe how experts regulate their thinking and thus avoid errors associated with cognitive overload (such as tunnel vision and goal fixation) during large-scale bushfires. The implications for bushfire training and procedures are discussed.     

Content and representation effects with reasoning tasks in PROLOG form

Abstract

Two experiments were carried out to examine human reasoning performance in the context of the logic programming language PROLOG. Two factors, ‘content’ (familiar versus unfamiliar) and ‘representation’ (diagrammatic versus PROLOG-like list) were investigated. Subjects answered questions about hierarchical relationships in each condition. A significant interaction was obtained in both experiments, subjects making fewer errors in the familiar-diagram and unfamiliar-list conditions than in the familiar-list and unfamiliar-diagram conditions.

It is hypothesized that a lower percentage of correct responses was given in familiar-list and unfamiliar–diagram conditions because the representation of information prevented successful use of an appropriate reasoning strategy. Working memory limitations provide a basis for understanding constraints on reasoning strategies for solving task questions. These strategies may involve either a serial or a spatial solution process. One strategy may require a larger working memory load than another, depending on the representation and content of the task information. Implications for PROLOG programming instruction are discussed.     

Private Camera Conversation: a new method for eliciting user responses?

Abstract

This paper discusses a new method called Private Camera Conversation. This method can be used to elicit user opinions on various subjects like their use of products. Initial impressions of the method are given on the basis of three small studies in which the method was employed. The method has not yet been validated and compared with existing human factors methods. This study reports the first findings of the application of the method. With Private Camera Conversation people are invited to talk about a particular topic in private to a video camera. The participants themselves decide when they want to start and when they want to stop the recording session. Initial impressions arc that the method has potential benefits for eliciting rich and useful responses even with respect to ‘personal’ issues. The Private Camera Conversation method seems particularly suitable for obtaining information about the social context in which products are used and about qualitative aspects of product use. The method is inviting and entertaining for participants and easy and effective to carry out. Optimal exploitation of the method has to be realized through further development. We hope this article will be a catalyst for further research.     

A laboratory study on preferred and imposed settings of a VDT workstation

Abstract

Thirty trained female typists performed several consecutive l0min typing tasks on an adjustable VDT workstation. The experiments were conducted with preferred as well as with imposed settings. The body postures were determined during the typing tasks and the subjects filled out questionnaires at the end of the tasks.

The study discloses that the preferred workstation dimensions are associated with mainly ‘relaxed’ sensations, while imposed dimensions—even if they are similar to the mean values of preferred settings—cause an increased incidence of static load symptoms in the sense of increased tension or impairments in the neck-shoulder-arm-hand area. The preferred keyboard levels are mainly distinctly higher than those recommended in standards and brochures. A chair with a high backrest and a proper support to rest forearms and wrists are preconditions for the preferred postures at VDT workstations.     

Using Neural Networks to predict HFACS unsafe acts from the pre-conditions of unsafe acts

Abstract

Human Factors Analysis and Classification System (HFACS) is based upon Reason’s organizational model of human error which suggests that there is a ‘one to many’ mapping of condition tokens (HFACS level 2 psychological precursors) to unsafe act tokens (HFACS level 1 error and violations). Using accident data derived from 523 military aircraft accidents, the relationship between HFACS level 2 preconditions and level 1 unsafe acts was modelled using an artificial neural network (NN). This allowed an empirical model to be developed congruent with the underlying theory of HFACS. The NN solution produced an average overall classification rate of ca. 74% for all unsafe acts from information derived from their level 2 preconditions. However, the correct classification rate was superior for decision- and skill-based errors, than for perceptual errors and violations.

Practitioner Summary: A model to predict unsafe acts (HFACS level 1) from their preconditions (HFACS level 2) was developed from the analysis of 523 military aircraft accidents using an artificial NN. The results could correctly predict approximately 74% of errors.     

Mining a new biometrics to improve the accuracy of keystroke dynamics-based authentication system on free-text

Related works for applying keystroke dynamics (KD) on free text identification indicated that applying KD can improve the accuracy of personal authentication on free text. As the result, this paper proposes a new biometrics, i.e., the keystroke clusters map (KC-Map), by clustering users’ keystrokes in order to effectively enhance the accuracy of personal authentication in free text. Since KC-Map is conducted via clustering, it is not suitable for traditional classifiers. In order to tackle this problem, the paper further proposes a keystroke clusters map similarity classifier (KCMS classifier). Experimental results positively show that the proposed KC-Map and KCMS classifier can efficiently improve the accuracy of personal authentication on free text with up to 1.27 times. In addition, one of the huge disadvantages on the current approaches in free text identification is that users are generally required to be trained for several months. Longer training time makes free text identification more impractical. Another motivation of this paper is to explore whether it is possible to shorten the training time into an acceptable range. Experimental results show that, to achieve relatively fair identification accuracy, users only need to carry out about 20 min for training.     

Variations in the Usability of Independent Web-Based Personal Health Records

ABSTRACT

Previous research on personal health records (PHRs) has focused on applications that are “tethered” to a specific electronic health record (EHR). However, there is a gap in research on the usability of unaffiliated, independent PHRs, as well as research on college-aged PHR users. Therefore, we performed a comparative usability study with 18 college-aged participants on three popular, freely available, independent PHRs. Participants completed the same six tasks in three different PHRs. Dependent variables included task time, mouse movement, mouse clicks, keystrokes, errors, and user satisfaction ratings based on the Computer System Usability Questionnaire (CSUQ). Analysis of variance (ANOVA) was used to determine the significance in the difference of the means for each dependent variable. Results showed statistically significant differences in CSUQ survey categories, errors, and keystrokes. Results supported one of the three PHRs (“System A”) as having better usability than its tested counterparts; participants scored best in System A of all categories of the CSUQ, and used System A with less mouse movement, commission errors, and keystrokes than the other PHRs.     

Speed and accuracy effects of fingers and dexterity in 5-choice reaction tasks

In two experiments we studied the influence of dexterity (controls, typists, pianists) and of differences between the five fingers of the dominant hand on speed and accuracy in a 5-choice reaction task. We used five coloured squares (Blue, Green, Yellow, Red, White) as stimuli, randomly varying foreperiods (3-10 s) and an intermediate stimulus-response-compatibility. The results reported here were independent of sex, foreperiods, and colour of stimuli.

In Experiment 1 (JV=168) with three groups (controls, typists, pianists) thumb and little finger showed significantly shorter reaction times than did index, middle, and ring finger. This difference did not interact with dexterity. Averaged across all the fingers typists and controls did not differ. Pianists showed significantly shorter reaction times than these two groups. Experiment 2 (N = 40) replicated all these findings

In both experiments groups did not differ in speed-accuracy-trade-off. The accuracy of the fingers was independent of dexterity. In all three groups the little finger showed the lowest rate of false alarms and the highest degree of reliability. The frequency distribution of finger confusions in all groups was in accordance with the spatial proximity of fingers: the closer their proximity the higher were the rates of confusion. In addition, these distributions showed an asymmetry across all three groups. The finger next to the 'correct’ finger in direction towards the thumb showed the highest false alarm rate in each case.     

Utilizing linguistically enhanced keystroke dynamics to predict typist cognition and demographics

Entering information on a computer keyboard is a ubiquitous mode of expression and communication. We investigate whether typing behavior is connected to two factors: the cognitive demands of a given task and the demographic features of the typist. We utilize features based on keystroke dynamics, stylometry, and “language production”, which are novel hybrid features that capture the dynamics of a typists linguistic choices. Our study takes advantage of a large data set (~350 subjects) made up of relatively short samples (~450 characters) of free text. Experiments show that these features can recognize the cognitive demands of task that an unseen typist is engaged in, and can classify his or her demographics with better than chance accuracy. We correctly distinguish High vs. Low cognitively demanding tasks with accuracy up to 72.39%. Detection of non-native speakers of English is achieved with F₁=0.462 over a baseline of 0.166, while detection of female typists reaches F₁=0.524 over a baseline of 0.442. Recognition of left-handed typists achieves F₁=0.223 over a baseline of 0.100. Further analyses reveal that novel relationships exist between language production as manifested through typing behavior, and both cognitive and demographic factors.     

A contextual study of police car telematics: the future of in-car information systems

The data from a former experiment with a serial responding task were analysed to throw light on some issues connected with the ‘ blocking ’ phenomenon. The task consisted of pushing one of four keys in response to the appearance of four figures on a numerical indicator. It was self-paced. Thirty-five subjects worked on it uninterruptedly for 30 min. It is shown that:

1. The main change which occurs in the distribution of reaction times during the session is the appearance of a tail of long times: this is in agreement with the blocking hypothesis, i.e. with the hypothesis that an extra delay is sometimes added to the normal reaction time.

2. The increase in number of long reaction times takes place in the first five minutes of work.

3. Long reaction times are preceded by an increase in mean reaction time and in percentage of errors, which take place over four or five preceding responses, and are followed by a return of both variables to normal level ; this is in agreement with the hypothesis that blocks allow fatigue to dissipate.     

Body composition and self-perceived health and fitness among indoor sports participants

Abstract

This paper examines (1) the responses of participants in specific indoor sports to questions regarding their self-perceived health and fitness, and (2) the relation betweeen these responses and body composition. A representative sample of indoor sports participants (n = 4308; 2938 male and 1370 female) from six cities in the United Kingdom responded to a questionnaire concerned with their self-perception of health and fitness. Height and weight was established from self-report and a body mass index (BMI) was calculated. Seventy eight percent of males and 85% of females reported their health as ‘good’ or ‘excellent’. Less than 2% of the sample reported their health as ‘poor’ or ‘very poor’. However, only 67% of males and 67% of females reported their fitness as ‘good’ or ‘excellent’. Again, the proportion of the sample reporting ‘poor’ or ‘very poor’ fitness was low; less than 4%. The question on health was compared with a recent general population survey and this yielded markedly more ‘unfavourable’ results; only 72% of males and 70% of females reported their health as being ‘good’ or ‘excellent’. A comparison of the response category groups for BMI showed significant differences between the ‘poor’ and ‘fair’ groups but not between the ‘good’ and ‘excellent’ groups.

It is concluded that: (1) indoor sports participants regard their health and fitness highly, but their fitness less so than their health; (2) adiposity tends to be associated with poorer perceptions of health and fitness; (3) women sports participants perceive their health more highly than others of the same age compared with men, yet the reverse is true of fitness.     

The effect of skill level,external frame of reference and environmental changes on one-handed catching

Abstract

Two groups of subjects designated ‘good’ (n= 12) and ‘poor’ (n= 10) catchers on the basis of a task-specific criterion were required to catch, one-handed, balls projected from a ball-projection machine under eight different conditions. Four of these comprised the ‘environment illuminated’ conditions (full light; hand obscured by an opaque screen; a see-through screen and an external vertical reference frame in the field of view) and four others the ‘environment dark’ conditions (catching hand wearing a glove painted with luminescent paint and u. v. illuminated ball in an otherwise dark room; ball illuminated by u. v. light; luminescent hand and external luminescent vertical reference plus u. v. illuminated ball; u. v.-illuminated ball, plus luminescent vertical reference).

While clear and significant differences - as was to be expected from the selection criteria imposed - were shown between the groups of ‘good’ and ‘poor’ catchers for all three dependent variables-number of catches, spatial errors and temporal errors-it was demonstrated that, for both groups, degrading the environment, by reducing the information available, resulted in a significant performance decrement on the dependent variables ‘number of catches’ and ‘temporal errors’. ‘Poor’ catchers were also shown to make significantly more spatial errors than ‘good’ catchers. Whereas the latter finding confirms earlier findings reported in the literature, the significant effect on performance of degrading the environment -particularly with respect to temporal errors - is new. The finding is discussed in the framework of the use of ‘tau’ (time to contact) in making judgements about when to initiate the grasp response.     

Task analysis for computer-aided design (CAD) at a keystroke level

The purpose of this research was to develop a new model to describe and predict a computerized task. AutoCAD was utilized as the experimental tool to collect operating procedure and time data at a keystroke level for a computer aided design (CAD) task. Six undergraduate students participated in the experiment. They were required to complete one simple and one complex engineering drawing. A model which characterized the task performance by software commands and predicted task execution time using keystroke-level model operators was proposed and applied to the analysis of the dialogue data. This task parameter model adopted software commands, e.g. LINE, OFFSET in AutoCAD, to describe the function of a task unit and used up to five parameters to indicate the number of keystrokes, chosen function for a command and ways of starting and ending a command. Each task unit in the task parameter model can be replaced by a number of primitive operators as in the keystroke level model to predict the task execution time. The observed task execution times of all task units were found to be highly correlated with the task execution times predicted by the keystroke level model. Therefore, the task parameter model was proved to be a usable analytical tool for evaluating the human-computer interface (HCI).     

On the discriminability of keystroke feature vectors used in fixed text keystroke authentication

Heterogeneous and aggregate vectors are the two widely used feature vectors in fixed text keystroke authentication. In this paper, we address the question “Which vectors, heterogeneous, aggregate, or a combination of both, are more discriminative and why?” We accomplish this in three ways - (1) by providing an intuitive example to illustrate how aggregation of features inherently reduces discriminability; (2) by formulating “discriminability” as a non-parametric estimate of Bhattacharya distance, we show theoretically that the discriminability of a heterogeneous vector is higher than an aggregate vector; and (3) by conducting user recognition experiments using a dataset containing keystrokes from 33 users typing a 32-character reference text, we empirically validate our theoretical analysis. To compare the discriminability of heterogeneous and aggregate vectors with different combinations of keystroke features, we conduct feature selection analysis using three methods: (1) ReliefF, (2) correlation based feature selection, and (3) consistency based feature selection. Results of feature selection analysis reinforce the findings of our theoretical analysis.     

Optimization of shift work: physiological contributions

Abstract

Healthy subjects (n= 62) with different circadian phase positions worked up to six weeks in different shift systems. Rotation speed was experimentally varied. On shift days all subjects worked in the Institute but spent their leisure time as usual.

In order to minimize bias through masking effects, ‘control days’ were interpolated (twice a week) into the shift periods. On these control days subjects remained on bedrest for 24 hours under constant conditions.

Results were as follows:

(i) The earlier the circadian phase position the poorer the adaptation to night-shifts. In particular, morning types were not able to adapt even to permanent night-shifts.

(ii) Evening types were able to adjust their circadian systems to a slowly rotating shift system (rotation speed seven days) and a 25 h-sliding system though they failed to do so with more quickly rotating systems. Whereas indifferent subjects were not able to adjust to even the slowly rotating system.

(iii) When stable phase shifts occurred in response to altered working hours they always amounted to about one hour per day if the working hours were not coherent with the usual environmental zeitgebers.

It is concluded that, on physiological criteria, very slow rotating shift systems or permanent night-shifts seem to be acceptable for evening types, but not for morning types who may require special health care in these conditions.     

Caractérisation de formations végétales mediterranéennes à partir de données ‘Thematic Mapper’ Une étude de cas en Andalousie (Espagne)

Abstract

The study is focused on the characterization of vegetation formations in a Mediterranean area (943 km²) located in southern Spain: herbaceous canopies (rangelands), shrubby vegetation (‘matorral’) and complex woody/herbaceous formations (‘dehesa’). Vegetation formations (physiognomical units) have been characterized by their spectral responses in the six reflective TM channels and by vegetation indices. From the ratio index TM4/TM3 there has been derived a map displaying seven classes (water, bare soil and five biomass levels reflecting the hierarchy of vegetation formations). Channels TM3, TM4 and TM5 have been considered for a supervised classification into nine land-cover categories (seven vegetation formations, bare soil and water). The proportion of correct classification of vegetation formations is about 78 per cent when considering test areas. Classification made from three principal components gives similar results.