ABSTRACT
Human face related digital technologies have been widely applied in various fields including face recognition based biometrics, facial landmarks based face deformation for gaming, facial reconstruction for those who are disfigured from an accident in the medical field and others. Such technologies typically rely on the information of a full, uncovered face and their performance would suffer varying degrees of deterioration according to the level of facial occlusion exhibited. 2D face recovery from occluded faces has therefore become an important research area as it is both crucial and desirable to attain full facial information before it is used in downstream tasks. In this paper, we address the problem of 2D face recovery from facial-mask occlusions, a pertinent issue that is widely observed in situations such as the Covid-19 pandemic. In recent trends, most researches are carried out through deep learning techniques to recover masked faces. The whole process consists of two tasks which are image segmentation and image inpainting. As U-Net is a typical deep learning model for image segmentation, but it also helpful in image inpainting and image colorization, so it has been frequently used in solving face recovery problems. To further explore the capability of U-Net and its variants for face recovery from masked faces, we propose to conduct a comparative study on several U-Net based models on a synthetic dataset that was generated based on public face datasets and mask generator. Results showed that Resnet U-Net and VGG16 U-Net had performed better in face recovery among the six different U-Net based models. © 2022 IEEE.
ABSTRACT
Objectives: Cardiac rehabilitation (CR) exercise classes in Singapore were either cancelled or suspended due to the ongoing COVID-19 pandemic. In response, a hybrid CR programme utilising telehealth was created to reduce reliance on in-person sessions. Our study aimed to assess early patients' impressions of this novel CR delivery method. Methods: Each hybrid CR programme comprises: (a) one CR orientation session via the Zoom video conferencing platform where patients interact with CR nurses and physiotherapists, (b) two 6-minute walk test fitness assessments (at baseline and upon completion), (c) five in-person supervised exercise classes and (d) two teleconsultation sessions (scheduled after exercise sessions 2 and 4) where CR physiotherapists review symptoms and discuss exercise prescriptions with patients over the phone. Upon programme completion, patients filled 2 anonymous online questionnaires - one assessing the videoconferencing component and another assessing the teleconsultation component and overall programme. Results: Seven patients underwent the hybrid CR programme. The majority were male (n=6, 86%), aged between 51 to 65 years (n=4, 57%) and had previously used a wide range of mobile applications (n=6, 86%). 100% of patients either agreed or strongly agreed that both videoconferencing and teleconsultation were: (a) easy to use, (b) as valuable as in-person sessions, (c) saved time and (d) had lower infection risk compared to in-person sessions (Figure). The overall ratings for videoconferencing, teleconsultation and the hybrid programme were 9.3, 9.6 and 9.6 (out of 10) respectively. All patients would definitely recommend the programme to others. Conclusion: Responses by participants of a novel hybrid CR programme utilising telehealth are highly encouraging. From a patient perspective, incorporation of telehealth to enable remote CR is feasible, acceptable and should be considered as an alternative CR delivery method.