1,721,023 research outputs found
Combining Forces: Data fusion across man and machine for biometric analysis
Full text linkThrough the HUMMINGBIRD framework outlined here,we seek to encourage a novel multidisciplinary approach to biometric analysis with the goal of enhancing both understanding and accuracy of identification
Expertise Effects in Voice Recognition
No full textResearch Objectives:Motivated by the other-race effect, the potential of an expertise effect in voice recognition is examined through manipulation of regional accent in a lineup study.Design:Identification of southern- and northern-accented voices by southern and northern participants was explored. Hits, misses and false alarms were recorded together with self-rated confidence.Method:49 Southerners and 85 Northerners took part in two lineups (southern voice, northern voice) each involving the presentation of a 60s target speaker clip, followed after a 5 minute delay by a 6 person voice lineup. Results:Northerners were good at identifying both the northern and southern voice. However, southerners showed a marked other-accent effect, with fewer hits and more false alarms to the northern voice than the southern voice.Conclusions:A parallel to the other-race effect emerged here for voice recognition. The implications of these findings are discussed with respect to eyewitness reliability.<br/
Introduction: giving each other a helping hand (in special issue on facial information processing: a multidisciplinary perspective)
No full text
Integration of a voice recognition route within IAC
No full textResearch Objectives: within an IAC framework, voice recognition and face recognition may usefully be viewed as parallel pathways. The demonstration of cross-modality priming of voices by faces and vice versa would add weight to such a view.Design: repetition priming of celebrity individuals was examined under within-modality and cross-modality conditions. Method: sixty participants completed a 2-stage method involving semantic decisions to celebrities at study, and speeded familiarity decisions to primed and unprimed celebrities at test.Results: expected within-modality priming effects emerged: Faces primed subsequent face recognition, and voices primed subsequent voice recognition. Additionally, voices successfully primed face recognition, however, faces did not prime subsequent voice recognition.Conclusions: the emergence of cross-modality priming adds support to the integration of voice recognition within a person perception model such as IAC. Results are discussed with respect to the relative strength of face and voice inputs in these two parallel pathways<br/
By which name should I call thee? The consequences of having multiple names
No full textThe nominal competitor effect suggests that, when a person has two names associated with them, recall of either name is more difficult than if they just had one name. Drawing on a connectionist framework, this effect could arise either if multiple names were represented as being connected to a single person identity node (PIN), or if multiple names were represented as being connected via one-to-one links to multiple PINs. Whilst the latter has intuitive appeal, results from two experiments support the former architecture. Having two names connected to a single PIN not only gives rise to a nominal competitor effect (Experiment 1), but also gives rise to a familiarity enhancement effect (Experiment 2). These empirical results are simulated using an extension of Brédart, Valentine, Calder, and Gassi's (1995) connectionist architecture, which reveals that both effects hold even when the association of both names to the PIN is unequal. These results are presented in terms of a more complete model for person recognition, and the representation of semantic information within such a model is examined
- …
