Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 9 de 9
Filter
1.
16th IEEE International Conference on Automatic Face and Gesture Recognition (FG) ; 2021.
Article in English | Web of Science | ID: covidwho-1853424

ABSTRACT

SARS-CoV-2 has presented direct and indirect challenges to the scientific community. One of the most prominent indirect challenges advents from the mandatory use of face masks in a large number of countries. Face recognition methods struggle to perform identity verification with similar accuracy on masked and unmasked individuals. It has been shown that the performance of these methods drops considerably in the presence of face masks, especially if the reference image is unmasked. We propose FocusFace, a multi-task architecture that uses contrastive learning to be able to accurately perform masked face recognition. The proposed architecture is designed to be trained from scratch or to work on top of state-of-the-art face recognition methods without sacrificing the capabilities of a existing models in conventional face recognition tasks. We also explore different approaches to design the contrastive learning module. Results are presented in terms of masked-masked (MM) and unmasked-masked (U-M) face verification performance. For both settings, the results are on par with published methods, but for M-M specifically, the proposed method was able to outperform all the solutions that it was compared to. We further show that when using our method on top of already existing methods the training computational costs decrease significantly while retaining similar performances. The implementation and the trained models are available at GitHub.

2.
20th International Conference of the Biometrics Special Interest Group, BIOSIG 2021 ; P-315:21-30, 2021.
Article in English | Scopus | ID: covidwho-1787337

ABSTRACT

The recent Covid-19 pandemic and the fact that wearing masks in public is now mandatory in several countries, created challenges in the use of face recognition systems (FRS). In this work, we address the challenge of masked face recognition (MFR) and focus on evaluating the verification performance in FRS when verifying masked vs unmasked faces compared to verifying only unmasked faces. We propose a methodology that combines the traditional triplet loss and the mean squared error (MSE) intending to improve the robustness of an MFR system in the masked-unmasked comparison mode. The results obtained by our proposed method show improvements in a detailed step-wise ablation study. The conducted study showed significant performance gains induced by our proposed training paradigm and modified triplet loss on two evaluation databases. © 2021 Gesellschaft fur Informatik (GI). All rights reserved.

3.
16th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2021 ; 2021.
Article in English | Scopus | ID: covidwho-1713989

ABSTRACT

SARS-CoV-2 has presented direct and indirect challenges to the scientific community. One of the most prominent indirect challenges advents from the mandatory use of face masks in a large number of countries. Face recognition methods struggle to perform identity verification with similar accuracy on masked and unmasked individuals. It has been shown that the performance of these methods drops considerably in the presence of face masks, especially if the reference image is unmasked. We propose FocusFace, a multi-task architecture that uses contrastive learning to be able to accurately perform masked face recognition. The proposed architecture is designed to be trained from scratch or to work on top of state-of-the-art face recognition methods without sacrificing the capabilities of a existing models in conventional face recognition tasks. We also explore different approaches to design the contrastive learning module. Results are presented in terms of masked-masked (M-M) and unmasked-masked (U-M) face verification performance. For both settings, the results are on par with published methods, but for M-M specifically, the proposed method was able to outperform all the solutions that it was compared to. We further show that when using our method on top of already existing methods the training computational costs decrease significantly while retaining similar performances. The implementation and the trained models are available at GitHub. © 2021 IEEE.

4.
20th International Conference of the Biometrics Special Interest Group, BIOSIG 2021 ; 2021.
Article in English | Scopus | ID: covidwho-1470286

ABSTRACT

The recent Covid-19 pandemic and the fact that wearing masks in public is now mandatory in several countries, created challenges in the use of face recognition systems (FRS). In this work, we address the challenge of masked face recognition (MFR) and focus on evaluating the verification performance in FRS when verifying masked vs unmasked faces compared to verifying only unmasked faces. We propose a methodology that combines the traditional triplet loss and the mean squared error (MSE) intending to improve the robustness of an MFR system in the masked-unmasked comparison mode. The results obtained by our proposed method show improvements in a detailed step-wise ablation study. The conducted study showed significant performance gains induced by our proposed training paradigm and modified triplet loss on two evaluation databases. © 2021 IEEE.

5.
20a Conferencia da Associacao Portuguesa de Sistemas de Informacao, CAPSI 2020 - 20th Conference of the Portuguese Association for Information Systems, CAPSI 2020 ; 2020-October, 2020.
Article in English | Scopus | ID: covidwho-1399989

ABSTRACT

Due to the pandemic crisis caused by COVID-19, people started performing their professional tasks remotely. This transition from the off-line and face-to-face work to online and remote work did not happen without pain, anxiety and some frustration as there was a need for accelerated learning and the frontiers between private and professional life were blurred. In this paper, we report on the preliminary findings of a survey-based research aimed at evaluating the main impacts and challenges faced by professionals in this transition process, in Portugal. Results show that respondents believe they work more, they feel more motivated, have better control over time, easily create methods and pace of work, and have more focus on tasks. Savings on fuel are offset by increased expenses in electricity, water, and food. Increased work-life balance but not increased leisure time are recognised. The most negative impact is social/professional seclusion, despite recognising fewer professional conflicts. © Atas da Conferencia da Associacao Portuguesa de Sistemas de Informacao 2020.

6.
World Conference on Information Systems and Technologies, WorldCIST 2021 ; 1367 AISC:81-91, 2021.
Article in English | Scopus | ID: covidwho-1265457

ABSTRACT

In Spring 2020, the world moved from traditional classes to what was coined as ERL (Emergency Remote Teaching/Learning/Instruction), posing real challenges to all actors involved, requiring an immediate, unprecedented, and unplanned devising of mitigation strategies. The impacts of this transition cannot, however, be studied only at the educational level, as it consists of a broader social shift with multidomain repercussions. In this paper, we use the CIPP model (Context, Input, Process and Product evaluations) to further investigate interrelations among the context, input and process elements of ERL during the first wave of COVID-19, as the second wave presses towards reconfining. A correlation analysis of 46 variables, based students’ responses (N = 360) to a closed-ended questionnaire shows the crucial importance of motivation and engagement in online classes, as learning enablers or constrainers. These also shape the students’ perception of the role that online classes play in helping them to stay more positive during ERL. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.

7.
World Conference on Information Systems and Technologies, WorldCIST 2021 ; 1366 AISC:34-40, 2021.
Article in English | Scopus | ID: covidwho-1265445

ABSTRACT

Traditionally, there have been many reasons and factors contributing to the disbelieve in remote work as a large-scale efficient work format. But with the social distancing imposed by the COVID-19 pandemic, remote work (RW) emerged as an adequate solution to continue labor at a large scale, even in the situations where remote work was unforeseeable and, in many cases, happened under inappropriate conditions. Despite some advantages, which are mainly identified by employees, there are other issues, pointed out by organizations, such as required technology infrastructure, HR training, etc. In any case, opting for remote work present some gains and losses for both employers and employees. With the end of the mandatory confinement, organizations reorganized the work format and, in some cases, it consisted in restoring the previous one. At this point, many questions arise being important to evaluate the perspectives of both sides – workers and companies – as well as their perceptions retrieved from the RW during the Spring confinement period, to understand how was the transition period and preview the future in terms of RW becoming part of the work model or even the way the pandemic situation is influencing future changes in businesses models. To achieve that goal, an exploratory survey-based research, and interviews will be carried out. The sample consists of the top 100 private companies of the Great Place To Work (GPTW) Portuguese ranking. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.

8.
Adv. Intell. Sys. Comput. ; 1328 AISC:303-314, 2021.
Article in English | Scopus | ID: covidwho-1204850

ABSTRACT

The pandemic caused by COVID-19 is not just a global crisis, it is ‘the first’ global crisis, and as the mandatory confinement shifted all education to Emergency Remote Instruction/Teaching/Learning, Higher Education Institutions were faced with the heavy task of balancing the immediate massive technological-pedagogical request by teachers and providing students with the socio-educational support that they need. This paper analyses the socio-educational impacts of the current confinement period on student’s lives and how they are responding to implemented ERL solutions, specifically in a stage of abundant pressing changes in which critical challenges are mostly felt. A close-ended questionnaire was built, comprising six dimensions of issues that may impact ERL: educational and organizational issues, technological and working conditions, social issues, family-related issues, psychological issues, and financial issues. Results were collected right after the first month of ERL and reveal that the most severe problems reside on the pedagogical and psychological domains. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.

9.
Smart Innov. Syst. Technol. ; 209:304-313, 2021.
Article in English | Scopus | ID: covidwho-968526

ABSTRACT

The recent pandemic caused by COVID-19 has forced unprecedented professional and social changes that affect the way people work, learn and live. Even in an era in which the digitalization of the workplace had been already widely announced, employers and employees had not expected such an urgent digital immersion of jobs with broad impacts on the professional, personal, financial and social dimensions of the workforce. In this paper, the main consequences, challenges, and effects of remote work in Portugal are reported, together with a set of recommendations for employers and workers, which were captured in survey-based research. Results show that professionals are not unanimous towards the benefits of remote work, and while some would like to resume in-person work, a substantial amount of professional, personal and financial benefits are associated to the possibility of working from home. Additionally, and according to pressing changes, professionals put forward several recommendations for employers and peers regarding workload and time management, employee assistance and self-regulated work. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

SELECTION OF CITATIONS
SEARCH DETAIL