Search for: All records

Numerous applications of Virtual Reality (VR) and Augmented Reality (AR) continue to emerge. However, many of the current mechanisms to provide input in those environments still require the user to perform actions (e.g., press a number of buttons, tilt a stick) that are not natural or intuitive. It would be desirable to enable users of 3D virtual environments to use natural hand gestures to interact with the environments. The implementation of a glove capable of tracking the movement and configuration of a user’s hand has been pursued by multiple groups in the past. One of the most recent approaches consists of tracking the motion of the hand and fingers using miniature sensor modules with magnetic and inertial sensors. Unfortunately, the limited quality of the signals from those sensors and the frequent deviation from the assumptions made in the design of their operations have prevented the implementation of a tracking glove able to achieve high performance and large-scale acceptance. This paper describes our development of a proof-of-concept glove that incorporates motion sensors and a signal processing algorithm designed to maintain high tracking performance even in locations that are challenging to these sensors, (e.g., where the geomagnetic field is distorted by nearby ferromagnetic objects). We describe the integration of the required components, the rationale and outline of the tracking algorithms and the virtual reality environment in which the tracking results drive the movements of the model of a hand. We also describe the protocol that will be used to evaluate the performance of the glove. 

Post-traumatic stress disorder (PTSD) is a mental health condition affecting people who experienced a traumatic event. In addition to the clinical diagnostic criteria for PTSD, behavioral changes in voice, language, facial expression and head movement may occur. In this paper, we demonstrate how a machine learning model trained on a general population with self-reported PTSD scores can be used to provide behavioral metrics that could enhance the accuracy of the clinical diagnosis with patients. Both datasets were collected from a clinical interview conducted by a virtual agent (SimSensei) [10]. The clinical data was recorded from PTSD patients, who were victims of sexual assault, undergoing a VR exposure therapy. A recurrent neural network was trained on verbal, visual and vocal features to recognize PTSD, according to self-reported PCL-C scores [4]. We then performed decision fusion to fuse three modalities to recognize PTSD in patients with a clinical diagnosis, achieving an F1-score of 0.85. Our analysis demonstrates that machine-based PTSD assessment with self-reported PTSD scores can generalize across different groups and be deployed to assist diagnosis of PTSD.