It is very characteristic of the perceptual system to test hypotheses and makes decision about the surrounding. This characteristic is a reflection of the interaction between outer object properties and the internal “orientation” factors when ocular fixations made at the object under focus. In other words, the connection between visual movements and sensory processing makes perception a passive experience and an active exploration of the surrounding. Scanning can reveal information critical to a certain perceptual decision. However, it is important to note observer factors can influence the scanning process. For instance, prior experience can influence the easiness and rapidity that it takes to reach a particular decision. This means that it will require an observer more fixation time to generate and process various sequences with novel items than when presented with familiar items (Barton, Radcliffe, Cherkasova, Edelman, " Intriligator, 2006).
Qualitatively, it is has been claimed that face processing ability varies from the any other pattern–processing abilities. According to Bredlau (2011) phenomenological theoretical view, the perception of faces varies from that objects because there always the likelihood of relational engagement. In particular, this very true for familiar faces, that is, faces of persons with a background of the actual relational engagement. The viewpoint and the familiarity of faces have been highlighted by several studies as important factors in face recognition among humans (Parr, Siebert, " Tauber, 2011). The viewpoint is the changes in the facial orientation from the vertical axis. Theoretically, it is argued inverting faces, that is, 180° face orientation can cause impaired recognition when processing both unfamiliar and familiar faces (Johnston " Edmonds, 2009). However, very few of these studies incorporate human factor such age and gender as possible covariates that may influence the cognitive process. It is very clear that age can affect cognitive abilities; for instance, aged persons sometime fail to recognize even close relatives.
The current study focuses on exploring differences in facial recognition reaction times (RTs) between inverted and upright stimulus. Familiarity will also be included as the primary covariate in exploring the effects on viewpoint on facial recognition RTs. In addition, the test will control for age and gender effects on the face recognition.
Methods
Design
The research study using an experimental design to test the paired samples, that is, the data was collected from one group. There independent variables which was either the face image orientation (upright or inverted) and the particular design of the independent variable was a within subjects design since the experiment was examining reactions time of two different stimuli from the same group of participants. There were two dependent variables being measured, which were whether participants were able to recognise the image of the face and whether they were able to correctly match the name to the face.
The design also controlled for extraneous variation by ensuring that all participants carried out their given tasks in the same environment and started the task at the same time hence minimizing any demand characteristics from the participants. Important extraneous variables such as age, gender and language capabilities were also included and recognised within the results.
Participants
Overall, the experimental study sampled data from 81 participants of which 56 were female and 25 were male. Years spent in the UK were also recorded in order to account for differences between scores from participants; the mean years spent in the UK was 17.8580. All participants were Birkbeck University psychology students and were recruited by the Birkbeck University of London. All participants were unaware of the tasks until they officially started the experiment and thus increases the validity of the results produced.
Materials
All participants were allocated to a computer where they will perform the given task via a link, participants were asked to read the consent form and were given the opportunity to either agree or disagree with the form (See Appendix A). Participants also had to fill a short form that was relevant to their demographical information including the age, gender, nationality and how many years they have resided in the UK (See appendix B). Before starting the initial task, participants were given clear instructions that they were required to pick only one answer that they deemed matched the image. The programme IBM SPSS statistical software was used to collect and analyse the raw data from the experiment.
Procedure
The experiment was computer-based where participants where provided with images and name choices to match up with recognized faces. After agreeing to the consent form and relevant information about themselves, upon the screen of the computer showed very clear instructions about how to complete the experiment (See appendix C). In total, the participant completed 40 trials each and within these trials, the participants were presented with either familiar (famous individuals) faces or unfamiliar faces (random faces). If participant’s response to the face was that it was familiar, they were given three options which were two names, one name is correct and the other is random and a “just familiar” option as well to identify the face (See appendix D). Some trails within the experiment included faces, which were upside down, but participants were told to ignore this and just answer normally. All data was then collected, cleaned, coded and analysed using SPSS. The data analysis includes both descriptive statistics and paired-sample t-tests to compare the mean RTs and mean percentage of correct name match ups between upright and inverted images.
Results
Descriptive statistics
Table I below indicates the summary statistics for the two dependent variables (performance and reaction time) grouped by the independent –viewpoint (inverted or upright). The performance measures the percentage of correctly matching faces with their names, where 80.12% and 61.30% of the upright and inverted faces respectively were correctly matched with their names. The mean RTs for upright images = 1770.63 milliseconds (ms) and mean RTs for inverted images =2225.36 ms and the standard deviation are 469.11 ms and 797.97 ms respectively.
Table 1: summary statistics for paired performances and reaction time
Mean
N
Std. Deviation
Std. Error Mean
Pair 1
Performance Upright
.8012
81
.12119
.01347
Performance Inverted
.6130
81
.10804
.01200
Pair 2
RTs Upright (ms)
1770.6296
81
469.10788
52.12310
RTs Inverted (ms)
2225.3642
81
797.98777
88.66531
Table 2 below shows the result of difference between each pair of the dependent variables. For the performance in matching names with faces, the test statistic t (80) =11.848, p-value < 0.05, which lead to rejection of the null hypothesis that the performance is the same for the inverted and upright face-name matching. For the RTs, the paired samples test statistic t (80) =-5.885, p-value < 0.05, which leads to the rejection of the null hypothesis that the reaction times for facial recognition does not differ between upright and inverted images.
Table 2: Paired samples t-test
Paired Differences
t
df
Sig. (2-tailed)
Mean
Std. Deviation
Std. Error Mean
95% Confidence Interval of the Difference
Lower
Upper
Pair 1
Performance (Upright – Inverted)
.18827
.14302
.01589
.15665
.21990
11.848
80
.000
Pair 2
RTs (Upright –Inverted)
-454.73457
695.46523
77.27391
-608.51456
-300.95458
-5.885
80
.000
Discussion and conclusion
The finding has indicated that there is a significant reaction time different when participant are required to differentiate inverted and upright faces. In particular, the difference indicates a negative difference, which implies that it takes more time to recognize inverted faces as compared to upright features. There is also statistically significant evidence that the performance is different between upright and inverted faces. Since the performance difference is positive, this implies that it is easier to match upright faces with names than inverted faces. In conclusion, viewpoint or orientation affects the face recognition in terms of reaction time and performance.
References
Barton, J. J., Radcliffe, N., Cherkasova, M. V., Edelman, J., " Intriligator, J. M. (2006). Information Processing during Face Recognition: The Effects of Familiarity, Inversion, and Morphing on Scanning Fixations. Perception, 35(8), 1089-1105. doi:10.1068/p5547
Bredlau, S. M. (2011). Monstrous faces and a world transformed: Merleau-Ponty, Dolezal, and the enactive approach on vision without inversion of the retinal image.Phenomenology and the Cognitive Sciences,10(4), 481-498. doi:10.1007/s11097-011-9210-6
Johnston, R. A., " Edmonds, A. J. (2009). Familiar and unfamiliar face recognition: A review. Memory, 17(5), 577-596.
Parr, L. A., Siebert, E., " Taubert, J. (2011). Effect of Familiarity and Viewpoint on Face Recognition in Chimpanzees.Perception,40(7), 863-872. doi:10.1068/p6971
