The limitations for expression recognition in computer vision introduced by facial masks.

Facial Expression recognition is a computer vision problem that took relevant benefit from the research in deep learning. Recent deep neural networks achieved superior results, demonstrating the feasibility of recognizing the expression of a user from a single picture or a video recording the face dynamics. Research studies reveal that the most discriminating portions of the face surfaces that contribute to the recognition of facial expressions are located on the mouth and the eyes. The restrictions for COVID pandemic reasons have also revealed that state-of-the-art solutions for the analysis of the face can severely fail due to the occlusions of using the facial masks. This study explores to what extend expression recognition can deal with occluded faces in presence of masks. To a fairer comparison, the analysis is performed in different occluded scenarios to effectively assess if the facial masks can really imply a decrease in the recognition accuracy. The experiments performed on two public datasets show that some famous top deep classifiers expose a significant reduction in accuracy in presence of masks up to half of the accuracy achieved in non-occluded conditions. Moreover, a relevant decrease in performance is also reported also in the case of occluded eyes but the overall drop in performance is not as severe as in presence of the facial masks, thus confirming that, like happens for face biometric recognition, occluded faces by facial mask still represent a challenging limitation for computer vision solutions.

Periocular Data Fusion for Age and Gender Classification.

In recent years, the study of soft biometrics has gained increasing interest in the security and business sectors. These characteristics provide limited biometric information about the individual; hence, it is possible to increase performance by combining numerous data sources to overcome the accuracy limitations of a single trait. In this research, we provide a study on the fusion of periocular features taken from pupils, fixations, and blinks to achieve a demographic classification, i.e., by age and gender. A data fusion approach is implemented for this purpose. To build a trust evaluation of the selected biometric traits, we first employ a concatenation scheme for fusion at the feature level and, at the score level, transformation and classifier-based score fusion approaches (e.g., weighted sum, weighted product, Bayesian rule, etc.). Data fusion enables improved performance and the synthesis of acquired information, as well as its secure storage and protection of the multi-biometric system's original biometric models. The combination of these soft biometrics characteristics combines flawlessly the need to protect individual privacy and to have a strong discriminatory element. The results are quite encouraging, with an age classification accuracy of 84.45% and a gender classification accuracy of 84.62%, respectively. The results obtained encourage the studies on periocular area to detect soft biometrics to be applied when the lower part of the face is not visible.

Telemedicine supported by Augmented Reality: an interactive guide for untrained people in performing an ECG test.

BACKGROUND: In many telemedicine applications, the correct use of medical device at the point of need is essential to provide an appropriate service. Some applications may require untrained people to interact with medical devices and patients: care delivery in transportation, military actions, home care and telemedicine training.Appropriate operation of medical device and correct connection with patient's body are crucial. In these scenarios, tailored applications of Augmented Reality can offer a valid support by guiding untrained people at the point of need. This study aims to explore the feasibility of using Augmented Reality in telemedicine applications, by facilitating an acceptable use of biomedical equipment by any unskilled person. In particular, a prototype system was built in order to estimate how untrained users, with limited or no knowledge, can effectively interact with an ECG device and properly placing ECG electrodes on patient's chest. METHODS: An Augmented Reality application was built to support untrained users in performing an ECG test. Simple markers attached to the ECG device and onto patient's thorax allow camera calibration. Once objects and their pose in the space are recognized, the video of the current scene is enriched, in real-time, with additional pointers, text boxes and audio that help the untrained operator to perform the appropriate sequence of operations. All the buttons, switches, ports of the ECG device together with the location of precordial leads were coded and indicated. Some user's voice commands were also included to improve usability. RESULTS: Ten untrained volunteers, supported by the augmented reality, were able to carry out a complete ECG test first on a mannequin and then on a real patient in a reasonable time (about 8 minutes on average). Average positioning errors of precordial electrodes resulted less than 3 mm for the mannequin and less than 7 mm for the real patient. These preliminary findings suggest the effectiveness of the developed application and the validity of clinical ECG recordings. CONCLUSION: This application can be adapted to support the use of other medical equipment as well as other telemedicine tasks and it could be performed with a Tablet or a Smartphone.