M2FRED: Mobile Masked Face REcognition Through Periocular Dynamics Analysis

Abstract

Recent regulations to block the widespread transmission of COVID-19 disease among people impose the use of facial masks indoor and outdoor. Such restriction becomes critical in all those scenarios where access controls take benefit from biometric recognition systems. The occlusions due to the presence of a facial mask make a significant portion of human faces unavailable for feature extraction and analysis. This work explores the contribution of the solely periocular region of the face to achieve a robust recognition approach suitable for mobile devices. Rather than working on a static analysis of the facial features, like largely done by work on periocular recognition in the literature, the proposed study focuses the attention on the analysis of face dynamics so that the spatio-temporal features make the recogniser frameindependent and tolerant to user movements during the acquisition. To obtain a lightweight processing, which is compliant with limited computing power of mobile devices, the spatio-temporal representation of the periocular region has analysed and classified through Machine Learning approaches. The experimental discussion has been performed on a new dataset, Mobile Masked Face REcognition Database, specifically designed to analyse the periocular region dynamics in presence of facial masks. For a wider comparative analysis, a publicly available dataset called XM2VTS has been considered as well as Deep Learning solutions have been experimented to discuss the challenging aspects of the recognition problem. Moreover, a summary of the state-of-the-art on periocular recognition driven by COVID pandemic has been presented, showing how the research efforts in this field focused on recognition of still images. Experimental results show promising levels of performance as well as limitations of the proposed approach, creating the premises for future directions.

M2FRED Analysis Using MobileNet and Siamese Neural Network

“…However, there is only a small number of research on mobile face recognition utilizing M2FRED. In [14], M2FRED was tested using MLP, SVM, DT, and RF, resulting in better performance compared to XM2VTS. Summary of the previous researches can be seen in Table I.…”

“…This research employed the Mobile Masked Recognition Through Periocular Dynamics Analysis (M2FRED) dataset [14], which consists of short videos of 43 subjects of various genders and ages uttering predetermined sentences or phrases. The videos were captured in an uncontrolled environment with four different conditions: indoors without a mask, indoors with a mask, outdoors without a mask, and outdoors with a mask as seen in Fig.…”

“…The best result for face recognition on the M2FRED dataset was achieved by using MLP with an accuracy of 95.4% [14].…”

“…Studies on mask detection and facemasked recognition are increasing rapidly to deal with this situation not only in algorithms but also in datasets. One of them was the Mobile Masked Face Recognition through Periocular Dynamics Analysis (M2FRED) dataset that was introduced by Cimmino et al [14] to study the periocular dynamic influence on face recognition.…”

Mobile Recognition of Image Components Based on Machine Learning Methods

“…However, some irresponsible people refuse to wear a mask despite all the dangers and consequences. Therefore, the development of a system, in particular using mobile technologies, to recognize the presence of a face mask in this case is very important [8][9][10]. Manually tracking the presence of a mask on people's faces is a difficult task and requires a lot of effort and additional staff.…”