Invisible touch—Control of a DICOM viewer with finger gestures using the Kinect depth camera L.C. Ebert a,n , G. Hatch b , M.J. Thali a , S. Ross c,d a University of Zurich, Institute of Forensic Medicine, Forensic Medicine and Imaging and Virtopsy, Winterthurerstrasse 190/52, CH-8057 Zurich, Switzerland b Centre of Forensic Imaging, Office of the Medical Investigator, University of New Mexico, MSC07 4040, 1101 Camino de Salud NE, Albuquerque, New Mexico 87102, USA c University of Bern, Institute of Forensic Medicine, Center for Forensic Imaging and Virtopsy, B¨ uhlstrasse 20, CH-3012 Bern, Switzerland d University of Link¨ oping, Center for Medical Image Science and Visualization (CMIV), SE-581 85 Link¨ oping, Sweden article info Article history: Received 17 July 2012 Received in revised form 25 October 2012 Accepted 5 November 2012 Available online 13 December 2012 Keywords: Virtopsy PACS OsiriX Kinect Gesture Detection abstract With the increasing use of imaging technologies during surgeries and autopsies, new control methods for computer systems are required to maintain sterility. Gesture controlled systems seem to be promising, since they allow for a touch-free control of computer systems. In a previous publication we presented a system which allows the control of the open source Picture Archiving and Communication System (PACS) OsiriX by means of gesture and voice commands. In order to overcome the limitations of this system, we developed a plug-in for OsiriX that allows for gesture control of the DICOM viewer of OsiriX with finger gestures. & 2012 Elsevier Ltd. All rights reserved. 1. Introduction Medical imaging datasets are increasingly used for planning and executing surgery [1] and autopsy [2]. This requires access to medical image viewers in the operating room and autopsy suite. However, using standard input devices such as mouse and key- board is a potential vector of infectious agents, posing risks to both patients and staff [3]. One solution to this problem is the use of gestures and voice commands to control imaging viewing software. Both are intuitive means of communicating and can allow for touch-free operation of computer systems. Systems that use standard color cameras for gesture identification have been presented, but they require calibration or can be sensitive to changes of lighting [4–6]. Several groups are working on systems combining image viewers with gesture control provided by the Kinect depth camera (Microsoft, Redmond, USA)[7–9]. Some commercial systems are in the making, such as the gesture control system TedCas from TedeSys (Spain)[10]. All these implementations use the skeleton tracking feature included in the Kinect programming interface. Since the skeleton tracking designed for gaming by tracking gross body movement, its accuracy is rather low. We endeavored to overcome this problem via the direct use of the depth images provided by the camera. Using the Kinect depth camera and a headset (Logitech ClearChat PC wireless headset, Logitech, Switzerland) for voice commands, we created and tested a working prototype of a software application that allowed users to control the open source PACS software OsiriX (OsiriX foundation, Geneva, Switzerland) by means of hand gestures and voice commands [11]. However, our study revealed some drawbacks. The voice recognition worked poorly if there was background noise or the user had a non-American accent. Also, the required headset was considered by many users to be ungainly and distracting. The rather simple method of gesture detection implemented in the first proto- type led to involuntary changes if the hands were moved in and out of the view volume. We concluded that more discrete detection of finger gestures might help to overcome these limitations. In this work, we present a new prototype that enables the user to control the OsiriX DICOM viewer with finger gestures using a low cost depth camera. It runs as a plug-in directly from OsiriX. 2. Materials and methods 2.1. Hardware For running the OsiriX as well as the Kinect plug-in, we used a MacBook pro (Intel Core i7 2.2 GHz, 8 GB RAM, Apple Inc., USA). Kinect A recent development in consumer electronics is the use of a low cost depth camera developed for gesture control of games. Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/jofri Journal of Forensic Radiology and Imaging 2212-4780/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jofri.2012.11.006 n Corresponding author. Tel.: þ41 44 635 6851. E-mail address: lars.ebert@virtopsy.com (L.C. Ebert). Journal of Forensic Radiology and Imaging 1 (2013) 10–14