Deep Learning Techniques Applied to Faces

Facial expressions are used by humans to convey their mood and emotional state to others. A listener grasps more efficiently speaker’s intentions and the content communicated to him/her, if he/she can observe speaker’s facial expressions, which complement speech prosody. Human facial expressions have a complex structure, requiring a good amount of time and practice to be decoded by the human brain, as well. Resorting to recent advancements in computer vision and machine learning, many researchers have promoted the state-of-the-art in facial expression recognition and have successfully exploited them in application domains, such as medical treatment, driver behavior analysis, and sociable robots deployment.

Earlier, researchers believed in that human facial expressions (e.g., the so-called six primitive emotions) remain invariant to cultural differences. But soon, it has been found out that there is no universal model of facial expressions and there are many deep facial expressions, which are affected by different factors.

Closely related to facial expression recognition, is understanding aging, which affects human face appearance in a progressive and accumulative manner. Although the effects of aging on human face are diverse and subject-dependent, some aging patterns, such as the formation of thin lines around eyes and/or mouth and the alteration of skin texture, are common across individuals. If these aging patterns are successfully learned, they can be exploited to simulate aging and generate subsequent face appearance (face age progression) or precedent face appearance (face age regression or rejuvenation). An efficient face aging method should produce realistic face aging results while maintaining personality, i.e., facial characteristics unique to everyone. Simulation of face age progression and regression is useful in security related applications, such as age invariant face recognition.

Recent advances in deep learning and artificial intelligence have contributed to understanding better facial expressions and advancing the performance of machine learning and/or synthesis tasks associated to face images. Deep learning basically tries to collect many small contributing features from a large human face image dataset (e.g., a dataset of human facial expressions). By doing so, it is ensured that even a small change can be recognized by the camera, which might get missed by human eyes also. Deep learning pipelines can be exploited in multiple application domains, such as biometrics, forensics, surveillance, medical treatment, social understanding, human-computer interaction. The correct recognition of feeling can enhance the security of biometric system. For example, a “feared face” can be declined to provide entry in a secured place even if face ID is correct.

The goal of this special issue is to gather highly novel contributions by prominent research scholars (either academicians industry professionals), which will shed light to deep learning techniques applied to faces and will advance related domain applications.

Topics are as below but are not limited to:

· Security issues solution with deep faces

· Deep faces as biometrics

· Facial expression and face aging based on generative adversarial networks

· Architectures, protocols, and algorithms for applications

· Smart medical solutions with deep face understanding

· Machine learning techniques for deep face recognition

· Artificial intelligence based face data generation and management

· High performance computing for deep faces

· Privacy concerns and solutions

· Cyber-physical systems and society

Schedule:

Deadline for submissions                                   September 15th, 2021

Notification of 1st revision round outcome       December15th, 2021

Revised submissions due to                                Feb 20th, 2022

Notification of acceptance                                   June 28h, 2022

Tentative publication date                                   2022

Potential Guest Editors:

1. Ioannis A. Kakadiaris

Director, CBL

Hugh Roy and Lillie Cranz Cullen University Professor

University of Houston

Houston, TX, USA

Email: Ikakadia@central.uh.edu

Webpage: https://uh.edu/cbl/people/about-director.php

2. Constantine Kotropoulos

Professor

Department of Informatics

Aristotle University of Thessaloniki

Thessaloniki, Greece

Email: costas@csd.auth.gr, costas@aiia.csd.auth.gr

Webpage: http://www.aiia.csd.auth.gr/LAB_PEOPLE/CKotropoulos.php

3. Assoc. Prof. Vitomir Struc, PhD
Laboratory for Machine Intelligence
Faculty of Electrical Engineering
University of Ljubljana
Ljubljana, Slovenia
Email: Vitomir.struc.fe.uni-lj.si

Webpage: https://lmi.fe.uni-lj.si/en/vitomir-struc/

4. Dr. Deepak Kumar Jain

Associate professor

State Key Laboratory of Pattern Recognition

Institute of Automation, Chongqing University of Posts and Telecommunications

Chongqing, China

Email: deepak@cqupt.edu.cn

Webpage: http://faculty.cqupt.edu.cn/DeepakKumarJain/zh_CN/index.htm