Master of Computing and Information Systems, Youngstown State University, 2010, Department of Computer Science and Information Systems

Face sketch recognition by humans has a significant value to both criminal investigators and researchers in computer vision, face biometrics, and cognitive psychology. An important question for both law enforcement agents and scientific researchers is how accurately humans identify hand-drawn face sketches correctly. However, the experimental studies of human performance in recognizing hand-drawn face sketches are still very limited in terms of the number of artists, the number of sketches, and the number of human evaluators involved. In this study, analysis has been concluded based on psychological tests in which 406 volunteers were asked to recognize 250 sketches drawn by 5 different artists. The primary findings are: i. The sketch quality has a significant effect on human performance. Inter-artist variation as measured by the mean recognition rate can be as high as 31%. ii. Participants showed a higher tendency to match multiple sketches to one photo than to second-guess their answers. The multi-match ratio seems correlated to recognition rate, while second-guessing had no significant effect on human performance. iii. For certain highly recognized faces, their rankings were very consistent using three measuring parameters: recognition rate, multi-match ratio, and second-guess ratio, suggesting that the three parameters could provide valuable information to quantify facial distinctiveness.

Committee: Yong Zhang PhD (Committee Co-Chair); John Sullins PhD (Committee Co-Chair); Graciela Perera PhD (Committee Member) Subjects: Computer Science; Information Systems; Psychological Tests; Psychology

Doctor of Psychology (Psy.D.), Xavier University, 2023, Psychology

Social anxiety involves excessive fear when there is a perceived threat of negative evaluation (Wong & Rapee, 2016). Individuals with social anxiety tend to use maladaptive strategies to avoid negative evaluation, such as experiential avoidance (Asher et al., 2021) and limiting intimate self-disclosure (Alden & Bieling, 1998). Recent data suggest that individuals with social anxiety tend to prefer online communication to face-to-face communication (Kamalou et al., 2019). Other studies suggest that individuals with social anxiety experience similar levels of anxiety during online and face-to-face communication (Doorley et al., 2020). It is unclear how past findings generalize to videoconference communication or face-to-face communication with a protective face mask. The purpose of the current study was to compare how individuals experience unmasked videoconference and masked face-to-face communication. Sixty-four undergraduate participants with traits of social anxiety engaged in a 45-minute conversation task. Dyads were randomly assigned to complete the task in a masked face-to-face or unmasked videoconference condition and completed measures of state anxiety, perceived threat of negative evaluation, experiential avoidance, and self-disclosure intimacy following the task. Analyses suggested the unmasked videoconference condition was associated with greater state anxiety and experiential avoidance, whereas the masked face-to-face condition was associated with greater perceived threat of negative evaluation. There was no difference between the two conditions on the self-disclosure measures. The results indicate that masked face-to face communication and unmasked videoconference communication have different effects on psychological processes associated with social anxiety.

Committee: Jennifer Phillips Ph.D. (Committee Member); Kathryn Smith Ph.D. (Committee Member); Nicholas Salsman Ph.D., ABPP (Committee Chair) Subjects: Behavioral Psychology; Behavioral Sciences; Clinical Psychology; Cognitive Therapy; Psychology

Ph.D., Antioch University, 2022, Antioch Seattle: Counselor Education & Supervision

The purpose of this research was to examine face-to-face and online supervisee disclosure within the supervisory alliance. Just as client care pivoted to online platforms, as a result of the COVID-19 pandemic, so did clinical supervision, which was uncharted territory for many, including those familiar with online counseling. The methodology used was consensual qualitative research (CSR). Eight participants were recruited as a sample of convenience, and semistructured interviews were conducted via Zoom. Results indicated domains such as important characteristics of the supervisory relationship, importance of communication, supervisor characteristics related to self-disclosure, positive aspects and negative aspects of online supervision, and positive and negative aspects of face-to-face supervision. In addition, relevant categories were identified. This study is relevant for counselors in training, counselor educators, and supervisors who are engaged in and considering online and face-to-face supervision. This dissertation is available in open access at AURA (https://aura.antioch.edu) and OhioLINK ETD Center (https://etd.ohiolink.edu).

Committee: Colin Ward PhD (Committee Chair); Mariaimeé Gonzalez PhD (Committee Member); Ned Farley PhD (Committee Member) Subjects: Behavioral Psychology; Clinical Psychology; Counseling Education; Counseling Psychology; Education; Psychology; Psychotherapy; Social Work

Master of Computer Science (M.C.S.), University of Dayton, 2021, Computer Science

Facial recognition with mask/noise has consistently been a challenging task in computer vision, which involves human wearing a facial mask. Masked Face Analysis via Multi-task deep learning is a method which will answer to many questions. In this thesis, we propose a unifying framework to simultaneously predict human age, gender, and emotions. This method is divided into three major steps; firstly, Creation of the dataset, Secondly, 3 individual classification models used for the system to learn the labelled (Age, Expression and Gender) images, Thirdly, the multi-task deep learning (MTDL) model; which takes the inputs as the data and shares their weight combined and gives the prediction of the person's (with mask) age, expression and gender. However, this novel framework will give better output then the existing methods

Committee: Tam Nguyen (Advisor); Ju Shen (Committee Member); Luan Nguyen (Committee Member) Subjects: Computer Science

Master of Science (MS), Ohio University, 2020, Computer Science (Engineering and Technology)

Automatically tracking and recognizing human faces in videos and live streams is often a crucial component in many high-level applications such as security, visual surveillance and human-computer interaction. Deep learning has recently revolutionized artificial intelligence areas, including face recognition and detection. Most of the existing video analysis solutions, however, rely on certain 2D convolutional neural network (CNN) to process video clips upon a frame-to-frame basis. The temporal contextual information between consecutive frames is often inadvertently overlooked, resulting in inconsistent tracking outcomes, which also negatively affect the accuracy of human identification. To provide a remedy, we propose a novel network framework that allows history information be carried along video frames. More specifically, we take the single short scale-invariant face detection (S3FD) as the baseline face detection network and combine it with long short-term memory (LSTM) components to integrate temporal context. Taking the images and detection results of previous frames as additional inputs, our S3FD + LSTM framework is well posed to produce more consistent and smoother face detection results along time, which in return leads to more robust and accurate face recognition in videos and live streams. We evaluated our face tracking and recognition model with both public (YouTube Face) and self-made datasets. Experimental results demonstrate that our S3FD+LSTM approach constantly produces smoother and more stable bounding boxes than S3FD alone. Recognition accuracy is also improved over the baseline model, and our model significantly outperforms the state-of-the-art face tracking solutions in the public domain.

Committee: Jundong Liu (Advisor) Subjects: Computer Science

Master of Science, The Ohio State University, 2020, Mechanical Engineering

In this study, computationally robust routines for hypoid surface generation and instantaneous contact line determination are proposed. The developed surface generation routine accurately defines undercut gear profiles for Face-Hobbed and Faced-Milled hypoid gears produced by Formate and generate processes. An approximation routine is incorporated into the developed surface generation model to obtain a stable initial guess for the non-linear solver. The proposed model successfully overcomes the computational challenges posed by higher order motion and complex blade profiles in the surface generation process. The efficacy of the developed routine was examined using industrial gear design having significant undercut, composite cutter profiles and generating motions. Unloaded Tooth Contact Analysis (UTCA) using the Ease-off method is outlined for obtaining the orientation and the location of the instantaneous contact lines. To enhance the accuracy for the prediction of the possible area of contact between gear and pinion, effects of roll angle on ease-off blank definition are investigated. A new blank definition routine for the Ease-off method is subsequently developed for accurate loaded contact analysis for the hypoid gears. The stabilization modules developed for the surface generation are extended to the Ease-off routines enhancing the robustness of the routines and enabling the visualization the complex cutter profiles (Toprem, Flankrem) and higher order motions.

Committee: David Talbot Dr. (Advisor) Subjects: Mechanical Engineering

Master of Arts, The Ohio State University, 2019, Psychology

The purpose of this study was to examine whether giving verbal description of the face can enhance the face recognition ability of individuals with Autism Spectrum Disorder (ASD). Participants in the current study were 18 individuals with ASD and 18 typically developing individuals (TD) with their full-scale IQ and age matched. The computerized experiment, which consisted of three experimental conditions, was conducted. Participants were exposed to three experimental conditions for a minute in a counterbalance order: 1) verbalization: giving verbal descriptions of the target faces 2) Navon: reporting global letters of the Navon stimuli 3) control: watching the video clip of a man riding a snowboard. The results showed a significant interaction effect between groups and conditions. Further analysis revealed that individuals with ASD reported higher accuracy in discriminating targets in the verbalization and Navon conditions than in the control condition. This result suggests that the face recognition ability of the individuals with ASD can be enhanced via verbal description of the face. Clinical implication and limitation of the study are further discussed.

Committee: Marc Tassé (Advisor); Luc Lecavalier (Committee Member); Susan Havercamp (Committee Member) Subjects: Psychology

MS, University of Cincinnati, 2018, Education, Criminal Justice, and Human Services: Information Technology

The purpose of this study is to analyze patterns of gender-based grade disparity in undergraduate information technology (IT) courses at a public research university. The study was conducted in an effort to understand factors in IT education that may contribute to the IT gender gap in higher education and careers. The study followed a learning analytics methodology developed and used by previous studies of gendered based performance differences in STEM courses at other universities. This research adds to a base of information on gender performance in IT courses. Results indicated that on average IT courses have gendered performance differences with the majority of courses favoring males. In addition, results found that on average course delivery method doesn't impact gender performance. A surprising result is that in the senior level courses, on average the courses favor females. The recommendations for expanding the study by examining additional course factors such as: (a) add additional course factors such as instructor gender, class size, and number of females in the course to determine the impact, (b) expand the data set to include additional computing disciplines, (c) expand the study to other universities. This study will lead to a research path to understanding gender neutral IT course instruction.

Committee: Hazem Said Ph.D. (Committee Chair); Chengcheng Li Ph.D. (Committee Member) Subjects: Information Technology

Master of Arts (M.A.), Xavier University, 2017, Psychology

The increasing use of information and communication technologies has allowed for a new type of incivility known as cyber incivility. This study investigated the incidence and outcomes of incivility committed face-to-face and online. Incivility has been associated with job satisfaction, burnout, turnover intentions, and neuroticism. This study compared the correlations of face-to-face and cyber incivility with the aforementioned organizational outcomes. Data were collected from 231 participants who completed an online survey asking about the aforementioned work outcomes and face-to-face and cyber incivility. Demographic information such as age, race, gender, job tenure, supervisory responsibilities, and average number of hours worked per week was also collected. Pearson Product-Moment correlations revealed that face-to-face and cyber incivility were both significantly correlated with each of the work outcomes consistent with previous research. Fisher's r-to-z transformations showed that face-to-face incivility was more strongly correlated with job satisfaction, burnout, turnover intentions, and neuroticism than cyber incivility. Both face-to-face and cyber incivility have negative implications for organizations and should be addressed.

Committee: Mark Nagy Ph.D (Committee Chair); Eric Barrett (Committee Member); Jeff Levey (Committee Member) Subjects: Psychology

Doctor of Philosophy (Ph.D.), University of Dayton, 2016, Educational Leadership

The purpose of this quantitative quasi-experimental study was to determine if students enrolled in an online introductory art appreciation course learned the same content as their fact-to-face counterparts. To achieve this goal, the researcher compared the level of knowledge attainment of course outcomes in four different content areas: the themes and purposes of art, the organizing principles of art, interpreting iconography, and differentiating between various art media (drawing, painting, sculpture). The following research questions guided the study: 1. How does the overall profile (gender, major, number of terms completed) of students enrolled in a face-to-face art appreciation course differ from that of students enrolled in an online art appreciation course?; 2. How does the gender profile of students enrolled in both the face-to-face and the online sections of an art appreciation course compare to that of students enrolled in courses in the greater academic division and the college?; 3. What impact does course format (face-to-face vs. online) of an introductory art appreciation course have on student achievement of course outcomes: familiarity with the themes and purposes of art, recognition of the organizing principles of art, ability to interpret iconography, and familiarity with various art media? The study's findings serve as an excellent point of departure for future research focusing on gender distribution in face-to-face art appreciation course sections, undeclared majors enrolled in online art appreciation course sections, and the impact that a variety of teaching styles and approaches to interaction may have on students' knowledge achievement of the art appreciation course outcomes.

Committee: Barbara M. De Luca Ph.D. (Committee Chair); C. Daniel Raisch Ph.D. (Committee Member); Peter J. Titlebaum Ed.D. (Committee Member); Mary A. Zahner Ph.D. (Committee Member) Subjects: Art Education; Art History; Community College Education; Education; Educational Leadership

Doctor of Philosophy, Miami University, 2013, Psychology

This paper investigates the role of category activation in the Own Group Bias in face recognition. In two experiments, I attempted to show that manipulations altering the salience of group memberships moderated the Own Group Bias, and that category activation mediated these effects. Experimental manipulations emphasized one of multiple possible group identities, and participants were expected to show the Own Group Bias along the salient category dimension. Experiment 1 was intended to establish the effectiveness of the manipulations in moderating the Own Group Bias, and Experiment 2 investigated the possible mediating role of category activation. Neither study supported the hypotheses, but Experiment 2 did provide some evidence for the moderation of the Cross Race Effect when other category dimensions are salient. Results are discussed with a focus on clarifying the conditions under which categorization may be most flexible and most likely to lead to differential Own Group Biases.

Committee: Kurt Hugenberg (Advisor); Amanda Diekman (Committee Member); Allen McConnell (Committee Member); Monica Schneider (Committee Member) Subjects: Social Psychology

Master of Computing and Information Systems, Youngstown State University, 2007, Department of Computer Science and Information Systems

Face recognition has received much attention recently in the biometrics research. Many studies have shown improvement in recognition rate when 2D and 3D faces were combined. However, the use of the 3D face has a few limitations such as that the 3D data requires much more storage space and long processing time. Therefore, there is a strong interest to explore new methods that can provide similar or better results in the face recognition. This thesis presents an experimental study by using a sequence of rotating head videos under two different lighting conditions, regular indoor lighting and strong shadow lighting. The experiment were carried out using two sets of data, the first set of over 100 subjects and the second set of 47 subjects. Very promising results have been observed in terms of the recognition performance measured by the cumulative characteristics curves.

Committee: Yong Zhang (Advisor) Subjects: Computer Science

Doctor of Philosophy, The Ohio State University, 2009, ED Teaching and Learning (Columbus campus)

This study compares the performance and experiences in an introductory statistics course across two modalities: web-based and classroom-based instruction. The research was conducted in classrooms from each learning environment for two teachers at a large midwestern community college. The results from the Pre-Test, Post-Test, and departmental final exam indicated that students in web-based statistics courses can have levels of average achievement comparable to that of their classroom-instructed counterparts. Yet, the facts that distance learners entered the course with stronger levels of mathematical preparedness, and had greater proportions of students at the extremes of the performance indicators, jointly challenge the notion of web-based instruction being “as good as traditional.” The faculty interviews, student questionnaires, and both virtual and physical instructional observations informed the researcher that the successful teaching strategies in the classroom can have an online comparative. Independent of modality, students desire teacher immediacy. It is also confirmed that the asynchronous nature of online learning that allows students to learn at their own pace will continue to drive student interest in spite of any potential barriers. The researcher recommends that future studies control for relevant student characteristics and any instructor effect to measure overall learning gains over longer periods of time. Hybrid courses were discussed as being the next modality on the horizon that would merit further research.

Committee: Patricia Brosnan PhD (Advisor); Dennis Pearl PhD (Committee Member); Terri Bucci PhD (Committee Member) Subjects: Mathematics Education

Doctor of Philosophy, The Ohio State University, 2007, Biomedical Engineering

Faces provide a wide range of information about a person's identity, race, sex, age, and emotional state. The perception of facial expressions of emotion is generally assumed to correspond to underlying muscle movement. However, the current work demonstrates that the static configuration of facial components biases the perception of the face. Furthermore, configural deviations from the population average influence the process of emotional attribution in the perception of neutral faces. Specifically, it is shown that changes in the relative position of the nose, mouth, eyes, and eyebrows affect the perception of emotional expression in an otherwise neutral face. The results of several experiments are presented wherein subjects were shown pairs of face images in which the relationship among the eyes, mouth, nose, and eyebrows was modified. Subjects were asked to respond by key-press to indicate their perception of a difference in emotional expression between the paired images. Results consistently showed that an increase in the distance between the eyes and mouth results in an increased perception of sadness. Similarly, a decrease in the distance between the eyes and mouth is perceived as an increase in anger. These perceptions occur in the absence of any changes in underlying facial musculature. The experiments confirm the notion that this perceptual bias is mediated by configural, rather than featural, processing. The results also support the contention that the center of the psychological face space for emotional expression represents the average of the faces most frequently experienced by the individual. A computational model is derived to simulate human perception of emotion in neutral faces.

Committee: Aleix Martinez (Advisor) Subjects:

Doctor of Philosophy, The Ohio State University, 2007, Physical Activity and Educational Services

Projections call for the proportion of students enrolled in blended courses to outpace the proportion of students in totally online courses in a few years. Yet the literature indicates that we know little about how undergraduates and graduate students learn in blended environments. In addition, few theoretical models have been developed specifically for online or blended learning. One such model the Community of Inquiry framework. The purpose of this study was to investigate how well the Community of Inquiry model explains discussion interactions among teaching presence, social presence, and cognitive presence over time and in different group modes to produce a worthwhile educational experience. Specifically, an undergraduate/graduate-level inquiry-based course at a large Midwestern university about the philosophical and historical roots of adult education in America was the context of the study. A quantitative content analysis of transcripts from one group (n = 4) that met via synchronous chats and another group (n = 5) that met face to face were examined. Statements in the transcripts were transformed to quantitative data for analysis within the framework of a mixed MANOVA design. Results found that group mode had an effect on frequency of social presence and cognitive presence. The face-to-face group generated more than twice the frequency of social presence indicators and nearly three times the frequency of cognitive presence indicators than the online group. There was no difference over time on the frequency of teaching presence, social presence, or cognitive presence for either the online or face-to-face group.

Committee: David Stein (Advisor) Subjects:

Doctor of Philosophy, The Ohio State University, 2003, Educational Theory and Practice

The use of electronic communication via networked computers has recently been evaluated as an effective medium for authentic communication in the field of foreign and second language instruction. This study investigated the effects of networked computers in foreign and second language learning by comparing electronic discussions (more specifically, synchronous online discussions) with traditional face-to-face discussions. The purpose of this study was to examine if there were any significant differences between electronic and traditional face-to-face communications in terms of linguistic productivity, syntactic complexity, linguistic accuracy and interactional features. The subjects of this study were ten NNSEs and ten NSEs, who were paired up to participate in the electronic discussion once and the face-to-face discussion once. The data elicited from the electronic and face-to-face discussions were examined and analyzed. The main frame of this study was a statistical analysis using paired t-tests and Wilcoxon matched-pairs signed-ranks tests. However, in order to better interpret findings obtained from this study, follow-up interviews were also conducted. The results of this study showed that electronic communication provided more opportunities for NNSEs to participate in the discussion and, as a result, the electronic mode contributed to more equal sharing of participation whereas face-to-face communication was more dominated by NSEs. In terms of syntactic complexity, the results exhibited that the discourse generated by electronic discussions contained a lower level of syntactic complexity than face-to-face discussions. With respect to grammatical accuracy, it was observed that NNSEs paid more attention to grammatical forms in the electronic discussions and made less grammatical errors under the electronic format. The findings of this study also showed that synchronous online discussion promoted interactive exchanges between communicators just as the traditional (open full item for complete abstract)

Committee: Charles Hancock (Advisor) Subjects:

Doctor of Philosophy, Miami University, 2011, Psychology

Two experiments tested the hypothesis that stereotypic expectancies exacerbate the cross-race effect in face recognition (CRE). Experiment 1 examined this stereotyping hypothesis by pairing cross-race (CR) Black targets and same-race (SR) White targets with either firearms or tools. Given the stereotypic linkage that exists between Blacks and firearms (Correll et al., 2002; Payne, 2001), the stereotyping hypothesis predicts an exacerbated CRE in the firearm relative to tool conditions. Furthermore, it was predicted that this increase should be predicted by the strength of participants‟ own race-weapon associations. Experiment 2 investigated attentional bias as a possible mechanism for these effects using a dot probe paradigm. Across both experiments, SR faces were better recognized than CR faces (consistent with the CRE), and faces were better recognized in the firearm than tool conditions. Neither the presence of a firearm nor the strength of participants‟ own race-weapon associations predicted the CRE. Possible theoretical implications for social cognitive models of the CRE and applied implications for eyewitness memory are discussed as well as possible directions for future research.

Committee: Kurt Hugenberg (Advisor); Heather M. Claypool (Committee Member); Maria L. Cronley (Committee Member); Amanda B. Diekman (Committee Member); Allen R. McConnell (Committee Member) Subjects: Psychology; Social Psychology

Master of Arts, Miami University, 2004, Psychology

The feasibility of using online psychotherapy process forms and the equivalence of these forms to their paper-and-pencil counterparts was investigated. Participants drawn from a psychology department clinic, a student counseling center, and community agencies visited an online Web site to report on each of their sessions by completing the two brief online rating forms. The feasibility of online psychotherapy process measure usage for therapists indicated they visited the Web site on a regular basis; however, client recruitment and participation was problematic. Equivalence between the online and paper-and-pencil forms was supported, with minor qualifications, by (i) similar underlying factor structures and (ii) comparable distributions of participant responses on the measures' scales. More direct contact with clients might increase their participation.

Committee: William Stiles (Advisor) Subjects: Psychology, Clinical

PHD, Kent State University, 2009, College of Education, Health, and Human Services / Department of Educational Foundations and Special Services

Four different data collection methods (face to face interviews, telephone interviews, mail surveys, and Web-based surveys) compared Social Desirability Bias (SDB). This study used Multiple Comparison tests and a randomized post-test only control-group design. No SDB differences were found among methods. For SDB no gender (2-factor), ethnicity (2-factor), nor was a gender and ethnicity (3-factor) interaction found. This study was more rigorous than other studies because 13 known extraneous influences were controlled for and two more were avoided. Effect sizes ranged from .002 to .029. Implications are: SDB need not trouble researchers when comparing the 4 methods and SDB may be decreasing over time. Suggestions for future research include (1) studies between non-published and published studies; (2) meta analytic method comparisons over well-established constructs; (3) meta analytic studies on SDB over time; (4) measurement invariance of the 4 methods on SDB; and other suggestions.

Committee: Rafa Kasim (Committee Co-Chair); Shawn Fitzgerald (Committee Co-Chair); Milton Harvey (Committee Member) Subjects: Social Research

Master of Science (M.S.), University of Dayton, 2012, Electrical Engineering

Automatic recognition of human faces (face photo recognition) irrespective of the expression variations and occlusions is a challenging problem. In the proposed technique, the edges of a face are identified, and a feature string is created from edge pixels. This forms a symbolic descriptor corresponding to the edge image referred to as 'edge-string'. The 'edge-strings' are then compared using the Smith-Waterman algorithm to match them. The class corresponding to each image is identified based on the number of string primitives that match. This method needs only a single training image per class. The proposed technique is also applicable to face sketch recognition. In face sketch recognition, a sketch drawn based on the descriptions of the victims or witnesses is compared against the photos in the mug shot database to facilitate a faster investigation. The effectiveness of the proposed method is compared with state-of-the-art algorithms on several databases. The method is observed to give promising results for both face photo recognition and face sketch recognition.

Committee: Vijayan K. Asari (Committee Chair); Tarek M. Taha (Committee Member); Eric J. Balster (Committee Member) Subjects: Computer Engineering; Computer Science; Electrical Engineering