A virtual system for postural stability assessment based on a TOF camera and a mirror Giuseppe Placidi giuseppe.placidi@univaq.it Naixia Pagnani naixia.pagnani@gmail.com Andrea Petracca andrea.petracca@graduate.univaq.it Matteo Spezialetti matteo.spezialetti@graduate.univaq.it A 2 VI_Lab, c/o Department of Life, Health and Environmental Sciences, University of L’Aquila Via Vetoio, 67100, L’Aquila, Italy Daniela Iacoviello iacoviello@dis.uniroma1.it Department of Computer, Control and Management Engineering Antonio Ruberti, Sapienza University of Rome Via Ariosto 25, 00185, Rome, Italy ABSTRACT Postural stability is often compromised in many pathological states and decreases with age. In clinical practice, an objective tool for balance is fundamental. Recently, virtual tools, based on the use of depth cameras, have been presented. In this paper, a new virtual system for postural stability assessment was presented, involving the use of a Time of Flight camera (TOF) and of a mirror for the reduction of the occlusions errors by allowing the camera to see the hidden body surface. The validity of the tool was assessed through some experimental results. Data were also compared with those measured by a physical force platform and those calculated with another virtual stability assessment system, in order to highlight the error reduction while maintaining simplicity and low-cost. CCS Concepts •Computing Methodologies ➝ Artificial Intelligence ➝ Computer Vision ➝ Image and video acquisition ➝ 3D imaging; •Computer Applications ➝ Life and medical sciences ➝ Health informatics. Keywords Balance Control; Postural Sway; COP; COG; Force Platform; TOF Camera; Mirror. 1. INTRODUCTION Balance control is the ability to maintaining the body Center Of Mass (COM) within its limits of stability. This capability, fundamental for controlling body movement, decreases with age [1] and could be compromised by many pathologies [2-4]. Both for diagnostic purposes and for assessing therapeutic progresses an objective and quantitative postural balance measure is needed. Recently, the effectiveness of a new generation of virtual instruments, exercises and practices for rehabilitation, have been studied and developed [5-9]. The assessment of the postural sway can be defined statically, if measurements are made while the subject tries to remain still standing, or dynamically, if the measurements are made under the effects of balance perturbations (these are important to assess the recovery from a loss of balance) [10]. Obviously, systems allowing dynamic measurements are also usable for static studies. Postural sway could be estimated starting from kinetic or kinematics parameters. Kinetic information include the excursion of the Center of Pressure (COP), applied to a support surface, and measured by means of clinical force platforms [11] or low-cost commercial instruments, like for example the Wii Balance Board [12]. Kinematic data could be used to estimate the spatial position of the Center of Mass and, consequently, its vertical projection on the ground, the Center of Gravity (COG). It could be measured by using wearable inertial sensors [13] or optical motion analysis, like that described in [14]. In particular, in [14] a low-cost tool for COM/COG assessment, based on a TOF camera was illustrated. During a virtual balance task, the COG excursions were recorded and compared with the movements done by COP, acquired by means of a force platform. Results showed that this tool was able to assess the sway of the human body also in dynamic conditions. The system had a lower dynamic range than a physical force platform, mainly due to the difference between COG and COP [15]. However, those differences were more evident in the Medio- Lateral (ML) direction of the subject movements than in the Antero-Posterior (AP) direction. This systematic error was produced because the Field Of View (FOV) of the camera was partial. Moreover, to ensure a real-time response, the model of the human body was approximated by a reduced set of spheres over the depth map. To overcome these issues, in this paper a refined version of the assessment method was presented. The method taken advantage from the presence of a mirror in the scene, whose orientation allowed the focusing of occluded portions of the body. The use of a mirror, instead of other TOF cameras, was twofold: it maintained low-cost; it avoided multiple-camera synchronization and high-frequency acquisition. The proposed method was validated by comparing the measured COM/COG movements with the COP excursions, observed by Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from Permissions@acm.org. REHAB '15, October 01-02, 2015, Lisbon, Portugal © 2015 ACM. ISBN 978-1-4503-3898-1/15/10…$15.00 DOI: http://dx.doi.org/10.1145/2838944.2838963