The present collection of studies examined the development of the ability to recognize facial identity rapidly and accurately, using two complementary approaches: comparing the performance of children and adults, and by training observers to learn novel stimuli in a laboratory setting. Across studies, children 8 to 10 years old performed less accurately than adults, a finding that confirms previous research that face processing takes many years to develop. However, results from two studies suggest that by 8 years of age, children encode individual facial identities relative to the average of previously experienced faces, in a manner similar to adults. The findings suggest that the basic mental architecture supporting face recognition is in place by 8 years. Additionally, children improved their ability to recognize unfamiliar faces from various viewpoints after just two, one-hour sessions of training, although the rate of learning was more variable than that observed in adults. The results from two studies also revealed that children's recognition accuracies of facial identity were lower than those of adults. An examination of children's similarity judgments of facial identity revealed that such immaturities in children's face processing may stem from greater variability in the mental representation of facial identities, rather than from immaturities in the encoding process per se. Findings from a final study suggest that the ability to make fine perceptual discriminations among individual faces arises, in part, from experience differentiating faces at the individual level, unlike the experience with non-face objects that typically involves recognition at the category level. The findings from the studies presented in this thesis suggest that such perceptual expertise may arise only with years of experience recognizing individual faces, and with sufficient neural development to support a stable mental representation of individual facial identities. / Thesis / Doctor of Philosophy (PhD)

Forged videos of swapped faces, so-called deepfakes, have gained a  lot  of  attention in recent years. Methods for automated detection of this type of manipulation are also seeing rapid progress in their development. The purpose of this thesis work is to evaluate the possibility and effectiveness of using deep embeddings from facial recognition networks as base for detection of such deepfakes. In addition, the thesis aims to answer whether or not the identity embeddings contain information that can be used for detection while analyzed over time and if it is suitable to include information about the person's head pose in this analysis. To answer these questions, three classifiers are created with the intent to answer one question each. Their performances are compared with each other and it is shown that identity embeddings are suitable as a basis for deepfake detection. Temporal analysis of the embeddings also seem effective, at least for deepfake methods that only work on a frame-by-frame basis. Including information about head poses in the videos is shown to not improve a classifier like this.

L’objectif de cette recherche est la création d’une plateforme en ligne qui permettrait d’examiner les différences individuelles de stratégies de traitement de l’information visuelle dans différentes tâches de catégorisation des visages. Le but d’une telle plateforme est de récolter des données de participants géographiquement dispersés et dont les habiletés en reconnaissance des visages sont variables. En effet, de nombreuses études ont montré qu’il existe de grande variabilité dans le spectre des habiletés à reconnaître les visages, allant de la prosopagnosie développementale (Susilo & Duchaine, 2013), un trouble de reconnaissance des visages en l’absence de lésion cérébrale, aux super-recognizers, des individus dont les habiletés en reconnaissance des visages sont au-dessus de la moyenne (Russell, Duchaine & Nakayama, 2009). Entre ces deux extrêmes, les habiletés en reconnaissance des visages dans la population normale varient. Afin de démontrer la faisabilité de la création d’une telle plateforme pour des individus d’habiletés très variables, nous avons adapté une tâche de reconnaissance de l’identité des visages de célébrités utilisant la méthode Bubbles (Gosselin & Schyns, 2001) et avons recruté 14 sujets contrôles et un sujet présentant une prosopagnosie développementale. Nous avons pu mettre en évidence l’importance des yeux et de la bouche dans l’identification des visages chez les sujets « normaux ». Les meilleurs participants semblent, au contraire, utiliser majoritairement le côté gauche du visage (l’œil gauche et le côté gauche de la bouche). / The present study aims to create a web-based platform that would examine individual differences in face processing strategies in different categorization tasks. The purpose of this platform is to collect data from geographically dispersed participants with variable face recognition abilities. Indeed, many studies have shown that there is high variability in the spectrum of face recognition ability, ranging from developmental prosopagnosia (Duchaine & Susilo, 2013), a disorder of face recognition in the absence of brain damage, to super-recognizers, individuals with extraordinary face recognition ability (Russell, Duchaine & Nakayama, 2009). Between these extremes, people vary substantially in their ability to recognize faces. To demonstrate the reliability of creating such a platform for individuals of widely varying abilities, we adapted a recognition task of the identity of famous faces using the Bubbles method (Gosselin & Schyns, 2001) and recruited 12 control subjects and a subject with developmental prosopagnosia. We were able to highlight the importance of the eyes and the mouth in face identification. The best observers seem mostly to use the left side of the face (left eye and the left side of the mouth).