Abstract
We analyze how people make similarity judgements amongst a variety of facial photos. Our investigation is based on the experiment conducted for 25 participants who were presented with a randomly ordered set of 356 photos (drawn equally from Caucasian and First Nations races), which they sorted based on their individual assessment. The number of piles made by the participants was not restricted. After sorting was complete, each participant was asked to label each of his or her piles with description of the pile’s content.
Race, along with labels such as “Ears” or “Lips”, may be treated as qualities that are a part of the judgement process used by participants. After choosing as many photos as possible, with the stipulation that half have the quality (called QP – for Quality Present) and half do not (QM – Quality Missing), we analyze the composition of each pile made by each participant. A pile is rated as QP (QM), if it contains significantly more of QP (QM) photos. Otherwise, it is QU (Quality Undecided). For the group of 25 participants, we form binary decision classes related to the QU percentages. For example, for the quality ”Ears”, a participant can drop into the class ”Uses-Ears”, if his or her QU is not more than, e.g., 60%, or the class ”Uses-Not-Ears” otherwise.
In our previous research, given the above decision classes, we constructed classifiers based on binary attributes corresponding to the judgements made by each participant about a specific pairs of photos; whether a participant rated them as similar (placed in same pile) or dissimilar (placed in different piles). Our goal was to find pairs indicating the quality occurrences in participants’ judgement processes. We employed rough set based classifiers in order to provide maximally clear interpretation of dependencies found in data. In this paper, we further validate those results by exploring alternative decision class compositions. For example, we find that manipulation of the QU threshold (fixed as 60% in our previous analyzes, as mentioned above) provides a useful parameter. This leads to an interaction between building a classifier and defining its training data – two stages that are usually independent from each other.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Bazan, J.G., Szczuka, M.: The rough set exploration system. In: Peters, J.F., Skowron, A. (eds.) Transactions on Rough Sets III. LNCS, vol. 3400, pp. 37–56. Springer, Heidelberg (2005)
Bergman, P., Berman-Barrett, S.J.: The Criminal Law Handbook: Know Your Rights, Survive the System, 11th edn., Nolo (2009)
Błaszczyński, J., Greco, S., Słowiński, R.: Multi-criteria classification – a new scheme for application of dominance-based decision rules. European Journal of Operational Research 181(3), 1030–1044 (2007)
Dysart, J., Lindsay, R., Hammond, R., Dupuis, P.: Mug shot exposure prior to lineup identification: Interference, transference, and commitment effects. Journal of Applied Psychology 86(6), 1280–1284 (2002)
Hepting, D., Maciag, T., Spring, R., Arbuthnott, K., Ślęzak, D.: A rough sets approach for personalized support of face recognition. In: Sakai, H., Chakraborty, M.K., Hassanien, A.E., Ślęzak, D., Zhu, W. (eds.) RSFDGrC 2009. LNCS, vol. 5908, pp. 201–208. Springer, Heidelberg (2009)
Hepting, D., Spring, R., Maciag, T., Arbuthnott, K., Ślęzak, D.: Classification of facial photograph sorting performance based on verbal descriptions. In: Szczuka, M. (ed.) RSCTC 2010. LNCS, vol. 6086, pp. 570–579. Springer, Heidelberg (2010)
Kelly, G.: The Psychology of Personal Constructs. Norton (1955)
Kimchi, R., Amishav, R.: Faces as perceptual wholes: The interplay between component and configural properties in face processing. Visual Cognition 18(7), 1034–1062 (2010)
Kuncheva, L.: Combining Pattern Classifiers: Methods and Algorithms. Wiley, Chichester (2004)
Pawlak, Z.: Rough set approach to knowledge-based decision support. European Journal of Operational Research 99, 48–57 (1997)
Pryke, S., Lindsay, R., Pozzulo, J.: Sorting mug shots: methodological issues. Applied Cognitive Psychology 14(1), 81–96 (2000)
Schooler, J.W., Ohlsson, S., Brooks, K.: Thoughts beyond words: when language overshadows insight. Journal of Experimental Psychology: General 122, 166–183 (1993)
Schwaninger, A., Lobmaier, J.S., Collishaw, S.M.: Role of featural and configural information in familiar and unfamiliar face recognition. In: Bülthoff, H.H., Lee, S.-W., Poggio, T.A., Wallraven, C. (eds.) BMCV 2002. LNCS, vol. 2525, pp. 643–650. Springer, Heidelberg (2002)
Ślęzak, D., Sosnowski, L.: SQL-based compound object comparators – a case study of images stored in ICE. In: Proc. of ASEA, Springer CCIS, vol. 117 (2010)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Hepting, D.H., Ślęzak, D., Spring, R. (2010). Validation of Classifiers for Facial Photograph Sorting Performance. In: Kim, Th., Pal, S.K., Grosky, W.I., Pissinou, N., Shih, T.K., Ślęzak, D. (eds) Signal Processing and Multimedia. MulGraB SIP 2010 2010. Communications in Computer and Information Science, vol 123. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-17641-8_23
Download citation
DOI: https://doi.org/10.1007/978-3-642-17641-8_23
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-17640-1
Online ISBN: 978-3-642-17641-8
eBook Packages: Computer ScienceComputer Science (R0)