Face perception

Early face-sensitive event-related potentials (ERPs) are modulated by age and race-of-face. Individual differences in implicit bias and race experience influence such race effects, but this remains largely unexplored. Thus, we examined children’s and adults’ P100s and N170s to own- and other-race faces and their relation with race experience and implicit racial bias. Children showed larger and more delayed P100s and N170s than adults. Also, 8- to 10-year-olds displayed earlier P100s to other-race faces, while 5- to 7- year-olds showed earlier left N170s to own-race faces. In adults, greater own-race experience was associated with delayed left N170s to own-race faces. Greater own-race bias was associated with earlier right P100s to own-race faces in 5- to 7-year-olds and smaller left P100 to N170 peak-to-peak amplitudes to own-race faces in 8- to 10-yearolds. Individual differences in age, race experience, and implicit racial bias should be considered when examining ERPs to own- and other-race faces.

Face perception and recognition abilities develop throughout childhood and differences in viewing own-race and other-race faces have been found in both children (Hu et al., 2014) and adults (Blais et al., 2008). In addition, implicit biases have been found in children as young as six (Baron & Banaji, 2006) and have been found to influence face recognition (Bernstein, Young, & Hugenberg, 2007). The current study aimed to understand how gaze behaviors, implicit biases, and other-race experience contribute to the other-race effect and their developmental effects. Caucasian children’s (5-10 years of age) and young adults’ scanning behaviors were recorded during an old/new recognition task using Asian and Caucasian faces. Participants also completed an Implicit Association Test (IAT) and a race experience questionnaire. Results found an own-race bias in both children and adults. Only adult’s IAT scores were significantly different from zero, indicating an implicit bias. Participants had a greater number of eye to eye fixations for Caucasian faces, in comparison to Asian faces and eye to eye fixations were greater in adults during encoding phases. Additionally, increased nose looking times were observed with age. Central attention to the nose may be indicative of a more holistic viewing strategy implemented by adults and older children. Participants spent longer looking at the mouth of Asian faces during encoding and test for older children and adults, but younger children spent longer looking at own-race mouths during recognition.
Correlations between scanning patterns and implicit biases, and experience difference scores were also observed. Both social and perceptual factors seem to influence looking behaviors for own- and other-race faces and are undergoing changes during childhood.

Third-order synthetic neural networks are applied to the recognition of isodensity facial images extracted from digitized grayscale facial images. A key property of neural networks is their ability to recognize invariances and extract essential parameters from complex high-dimensional data. In pattern recognition an input image must be recognized regardless of its position, size, and angular orientation. In order to achieve this, the neural network needs to learn the relationships between the input pixels. Pattern recognition requires the nonlinear subdivision of the pattern space into subsets representing the objects to be identified. Single-layer neural networks can only perform linear discrimination. However, multilayer first-order networks and high-order neural networks can both achieve this. The most significant advantage of a higher-order net over a traditional multilayer perceptron is that invariances to 2-dimensional geometric transformations can be incorporated into the network and need not be learned through prolonged training with an extensive family of exemplars. It is shown that a third-order network can be used to achieve translation-, scale-, and rotation-invariant recognition with a significant reduction in training time over other neural net paradigms such as the multilayer perceptron. A model based on an enhanced version of the Widrow-Hoff training algorithm and a new momentum paradigm are introduced and applied to the complex problem of human face recognition under varying facial expressions. Arguments for the use of isodensity information in the recognition algorithm are put forth and it is shown how the technique of coarse-coding is applied to reduce the memory required for computer simulations. The combination of isodensity information and neural networks for image recognition is described and its merits over other image recognition methods are explained. It is shown that isodensity information coupled with the use of an "adaptive threshold strategy" (ATS) yields a system that is relatively impervious to image contrast noise. The new momentum paradigm produces much faster convergence rates than ordinary momentum and renders the network behaviour independent of its training parameters over a broad range of parameter values.

Two experiments were conducted to investigate the impact of poser and perceiver gender on the Happiness/Anger Superiority effect and the Female Advantage in facial expression recognition. Happy, neutral, and angry facial expressions were presented on male and female faces under Continuous Flash Suppression (CFS). Participants of both genders indicated when the presented faces broke through the suppression. In the second experiment, angry and happy expressions were reduced to 50% intensity. At full intensity, there was no difference in the reaction time for female neutral and angry faces, but male faces showed a difference in detection between all expressions. Across experiments, male faces were detected later than female faces for all facial expressions. Happiness was generally detected faster than anger, except when on female faces at 50% intensity. No main effect for perceiver gender emerged. It was concluded that happiness is superior to anger in CFS, and that poser gender affects facial expression recognition.

Many errors in recognition are made because various features of a stimulus are attended inefficiently. Those features are not bound together and can then be confused with other information. One of the most common types of these errors is conjunction errors. These happen when mismatched features of memories are combined to form a composite memory. This study tests how likely conjunction errors, along with other recognition errors, occur when participants watch videos of people both with and without unusual facial features performing actions after a week time lag. It was hypothesized that participants would falsely recognize actresses in the conjunction item condition over the other conditions. The likelihood of falsely recognizing a new person increased when presented with a feature, but the conjunction items overall were most often falsely recognized.

This dissertation introduces our work on face recognition using a novel approach based on creating 3D face model from 2D face images. Together with the pose angle estimation and illumination compensation, this method can be used successfully to recognize 2D faces with 3D recognition algorithms. The results reported here were obtained partially with our own face image database, which had 2D and 3D face images of 50 subjects, with 9 different pose angles. It is shown that by applying even the simple PCA algorithm, this new approach can yield successful recognition rates using 2D probing images and 3D gallery images. The insight gained from the 2D/3D face recognition study was also extended to the case of involving 2D probing and 2D gallery images, which offers a more flexible approach since it is much easier and practical to acquire 2D photos for recognition. To test the effectiveness of the proposed approach, the public AT&T face database, which had 2D only face photos of 40 subjects, with 10 different images each, was utilized in the experimental study. The results from this investigation show that with our approach, the 3D recognition algorithm can be successfully applied to 2D only images. The performance of the proposed approach was further compared with some of the existing face recognition techniques. Studies on imperfect conditions such as domain and pose/illumination variations were also carried out. Additionally, the performance of the algorithms on noisy photos was evaluated. Pros and cons of the proposed face recognition technique along with suggestions for future studies are also given in the dissertation.

This study aimed to understand the differences in strength or coordination of brain regions involved in processing faces in the presence of aging and/or progressing neuropathology (Alzheimer's disease). To this end, Experiment 1 evaluated age-related differences in basic face processing and the effects of familiarity in face processing. Overall, face processing in younger (22-35yrs) and older participants (63-83yrs) recruited a broadly distributed network of brain activity, but the distribution of activity varied depending on the age of the individual. The younger population utilized regions of the occipitotemporal, medial frontal and posterior parietal cortices while the older population recruited a concentrated occipitotemporal network. The younger participants were also sensitive to the type of face presented, as Novel faces were associated with greater mean BOLD activity than either the Famous or Relatives faces. Interestingly, Relatives faces were associated with greater mean B OLD activity in more regions of the brain than found in any other analysis in Exp. 1, spanning the inferior frontal, medial temporal and inferior parietal cortices. In contrast, the older adults were not sensitive to the type of face presented, which could reflect a difference in cognitive strategies used by the older population when presented with this type of face stimuli. Experiment 2 evaluated face processing, familiarity in face processing and also emphasized the interactive roles autobiographical processing and memory recency play in processing familiar faces in mature adults (MA; 45-55yrs), older adults (OA; 70-92yrs) and patients suffering from Alzheimer's disease (AD; 70-92yrs).

This study provides evidence for an age bias in face recognition. Younger adults viewed short video clips of young actors or of actors over the age of 60 performing everyday actions. One week later, participants were tested on their memory for these events. Recognition event types included same, completely new, and conjunction items. In conjunction items, a familiar actor performed a familiar action that had actually been performed by someone else during encoding. Participants performed well at picking out the new and old events, but had more difficulty distinguishing between the conjunction events. Younger adults were significantly worse at recognizing the conjunction items when the age of the actor was different from encoding to retrieval. This study supports the hypothesis that people are better able to recognize and distinguish others within a similar age range compared to people outside that range.

Event memory studies have shown that older adults typically have poorer memories for events than do younger adults. Unfortunately, these studies tested memory for events that contained only young adults as the actors. It could be that the younger adults remembered the events better due to an own-age bias. One particular type of event memory error is called unconscious transference. In unconscious transference, a binding error results in false conjunction memories. Several studies have shown that older adults are more likely to make these errors. This study aims to look more closely at the effects of own-age bias on face recognition and conjunction errors made in memory for events. Younger adults were tested on their memory for both younger and older adults seen performing simple actions in video clips as well as in "mug shot" photograph tests.