Intelligent Data Analysis: An Introduction
Intelligent Data Analysis: An Introduction
Combining Pattern Classifiers: Methods and Algorithms
Combining Pattern Classifiers: Methods and Algorithms
Ethics and usability testing in computer science education
ACM SIGCSE Bulletin
A Rough Sets Approach for Personalized Support of Face Recognition
RSFDGrC '09 Proceedings of the 12th International Conference on Rough Sets, Fuzzy Sets, Data Mining and Granular Computing
Classification of facial photograph sorting performance based on verbal descriptions
RSCTC'10 Proceedings of the 7th international conference on Rough sets and current trends in computing
The rough set exploration system
Transactions on Rough Sets III
| Hi-index | 0.00 |
Facial recognition is routine for most people; yet describing the process of recognition, or describing a face to be recognized reveals a great deal of complexity inherent in the activity. Eyewitness identification remains an important element in judicial proceedings: it is very convincing, yet it is not very accurate. A study was conducted in which participants were asked to sort a collection of facial photographs into an unrestricted number of piles, based on their individual judgements of facial similarity. Participants then labelled each pile. Three different qualities identified in the photos may have been used by participants in making similarity judgements. Choosing photos with the stipulation that half have the quality and half do not, the composition of each pile made by each participant was analysed. A pile is rated as "quality present" (or missing), if it contained significantly more of "quality present" (or missing) photos. Otherwise, it was rated as "quality undecided". As a result, an information system was obtained with objects corresponding to the participants and attributes corresponding to the pairs of photos. Further, a decision attribute was added for each of the qualities. The decision classes were determined by setting a threshold for the percentage of QU photos. Initially, this threshold was determined by observation. The rough set based attribute reduction methodology was applied to this data in order to build classifiers for each quality. Other thresholds for QU were also considered, based on computational experimentation, in order to improve the accuracy of the classifiers. This paper describes the initial study, the computational approach which includes an important pre-processing step and development of a useful heuristic, the results from the evaluation, and a list of opportunities for future work. Although different participants created quite different sortings of photos, the rough set analysis detected photos with more general significance. This may lead to a practical test for individual abilities, as well as to inferring what discriminations people use in face recognition.