You are on page 1of 5

Mission Impossible Extended Room

Jordan van der Kroon 05914732 8dkjv@queensu.ca Khurrum Abdul Mujeeb 05801597 7mka2@queensu.ca Damilare Awosanya 10073810 12da15@queensu.ca

Abstract
Our goal was to recreate the Kremlin hallway scene from the movie Mission Impossible: Ghost Protocol. To accomplish this we used the big screen in our lab as a window to look into a 3D virtual extension of the lab. We created a 3D virtual space that mirrored the real lab as well as including a natural extension of the lab. The goal here is to use the big screen as a window into the virtual space to delude a user to think the lab is extended past the screen when they look at it. To accomplish this we would create a window into this virtual reality for the user to peer into on the x,y plane that would represent the big screen in our 3D virtual space. We can then map the user’s position in the virtual lab using a Kinect and using an avatar representation of the user’s position in the real lab. The image on screen is dynamically updated by drawing a line from the avatar’s head through the big screen to determine the camera position. This line goes through the screen towards the middle of the back wall of the extended lab to determine the camera rotation.

Keywords
Blended Reality, Avatar, Mission Impossible Ghost Protocol, Skeleton Tracking, Big screen, Microsoft Kinect
Copyright is held by the author/owner(s). CISC425, March 25, 2013, Kingston, Ontario, Canada.

ACM Classification Keywords
H.5.2 [Information interfaces and presentation]: User Interfaces.

2

Introduction
In the movie, the representation of the hallway projected on the big screen was accurate enough to fool the security guard at the opposite end of the hallway into thinking he was looking down the length of the hallway normally. In the movie the screen uses eye tracking and a projector mounted on a moveable robotic arm to accomplish this. Once a second security guard enters in front of the screen it fails to lock onto a single user and the image projected on screen appears to be bouncing back and forth between the two perspectives. Our goal for this project was to show a natural extension of the lab on screen that dynamically updated the perspective based on a single user’s position in front of the screen. The image shown on screen would accurately match what a user would see if they were in the virtual 3D space we had created. Therefore we were required to build a 3D virtual space that would accurately represent the real lab and a user’s position inside it. The next step was to add an extension to this space where the big screen is normally. This extension is what would be shown on screen so as to delude the user that what they are looking at is a lab that is twice as big as it really is.

wanted to eliminate the mapping of physical input on a user interface to interactions in the virtual world. To accomplish this they use the screen as a window into their virtual world where the user can interact with the virtual objects without the use of a representation of the user or an avatar to do so. They accomplish this in their 3D model of the “Apple Yard” by having apples fly out of the virtual world through the window that is their screen and have users hit them using a hand held wand. Our implementation is similar to this except for the fact that we have a representation of the real world in our virtual world. This representation is not in view to the user because our window into the virtual lab sits on the xy plane where the representation of the real lab ends and our extended virtual lab begins. Our avatar that represents the user and their position in the lab is behind this window which is the only perspective the user has into the virtual world. The window through which the user can see into the virtual extended lab is updated based on the user’s position in front of the screen. The avatar mirrors the player’s movement and position around the real lab and the perspective shown on screen is dynamically updated to represent the user’s new position.

Related Works
Blended Reality To make the user feel like they are looking through the screen into the natural extension of the lab we wanted to use the experience of “Blended Reality” (BR). The idea of the big screen being a window of perspective into the virtual room is based on the research done by Huynh[6]. In their work of blended Reality they

Implementation 3D Model
We had access to the 3D model that was used as a blue print when building the lab. The model we received had no textures or lighting which caused everything in the model appeared black. This made it very difficult to identify key parts of the lab. A light source and white textures were added to most of the components that

3

made up the lab to be able to correctly identify them. We were only concerned with a single room of the lab. We removed everything that was not incorporated in this room.

Figure 1. Unmodified Model with white textures added

the other side of the screen. We did want to maintain the design by keeping a row of pods on the right side and a large sprawling desk on the left. The 3D virtual lab we created is asymmetrical. The concept behind that came from the design on the windows. The window behind the big screen in the real world is actually a perfect cross while all the windows to the left (when looking outside of the lab) of it start skewing the cross to the left. The symmetrical point of the extended lab is the perfect cross behind the big screen. All of the windows to the right are skewing to the right in the same style as the real lab.

Figure 2. Trimmed down model with textures added from High-Res photos taken from inside the HML. Big screen can be seen here in black.

Figure 3. Asymmetrical extended lab windows

Our goal here was to create an extended version of our lab with the big screen to be the point where the real lab ends and the extended lab can be seen. To track the position of the user in the real lab we would have an avatar representation of the user to actually move around the real lab side of the 3D virtual world. With the position of this avatar in the real side of the virtual lab we would be able to show the perspective of the avatar looking at through the big screen. To create a natural extension of the lab we wanted to incorporate similar elements in the extension that were inside the real lab. To accomplish the natural feel we didn’t want to simply show a mirror image of the lab on

Figure 4. Bird’s eye view of asymmetrical extended lab(left) next to the real lab(right)

Kinect Implementation
In order to extrapolate real-world data to use as an input device to move around the extended Lab, the use of a single Kinect and the Kinect SDK 1.6 was implemented. Since we are using unity to render the extended lab model, the Kinect SDK alone did not suffice. In order to map data from the Kinect onto unity, we required a Kinect to unity wrapper, which was

4

available in the form of Zifgu – an open source Kinectunity wrapper. Along with the ability to map data from the Kinect to unity, Zigfu provides useful sample scripts. The ‘Blockman3rdPerson’ sample script was used as the bases for our project. The blockman container in the script simulates a user skeleton and by placing the blockman container onto the extended lab model we were able to simulate the location of the person inside the real lab. Once the blockman container was placed onto the extended lab with our specifications, transparency was applied to the container to provide a first-person view of the lab. Adjustments to the camera were made to ensure reallife movement was replicated onto the extended lab, the camera was moved from behind the head of the blockman container to in front of it as seen in Figure 6. Here the camera has a rooted z coordinate. Which means it is only free to move in the xy plane that represents where the big screen exists in the lab shown by the black line in figure 5-8. The camera is then targeted to always be facing the far back wall of the extended lab. A line is drawn from blockman’s head through the xy plane that is the big screen all the way to the back wall of the extended lab. Camera position is determined by where that line that intersects the xy plane which is shown by the yellow ring in Figure 5 and Figure 7. Camera rotation is determined by following the line to the center of the back wall where the camera is always updated to point at.

Figure 5. The yellow circle is where the camera is situated on the xy plane. The red line then follows the camera direction into the back wall.

Figure 6. View of the camera positioned where the yellow circle was in figure 5 pointed at the back wall.

5

2.

Kibum Kim, John Bolton, Audrey Girouard, Jeremy Cooperstock, and Roel Vertegaal. 2012. TeleHuman: effects of 3d perspective on gaze and pose estimation with a life-size cylindrical telepresence pod. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '12). ACM, New York, NY, USA, 2531-2540. DOI=10.1145/2207676.2208640 Dmitry Batenkov. 2010. Real-time detection with webcam. XRDS 16, 4 (June 2010), 50-51. DOI=10.1145/1764848.1764861 Jiří F. Urbánek, Theodor Baláž, Jiří Barta, and Jaroslav Prücha. 2011. Technology of computeraided adaptive camouflage. In Proceedings of the 2011 international conference on Computers and computing (ICCC'11), Vladimir Vasek, Yuriy Shmaliy, Denis Trcek, Nobuhiko P. Kobayashi, and Ryszard S. Choras (Eds.). World Scientific and Engineering Academy and Society (WSEAS), Stevens Point, Wisconsin, USA, 81-87. Hao Du, Peter Henry, Xiaofeng Ren, Marvin Cheng, Dan B. Goldman, Steven M. Seitz, and Dieter Fox. 2011. Interactive 3D modeling of indoor environments with a consumer depth camera. In Proceedings of the 13th international conference on Ubiquitous computing (UbiComp '11). ACM, New York, NY, USA, 75-84. DOI=10.1145/2030112.2030123 David F. Huynh, Yan Xu, and Shuo Wang. 2006. Exploring user experience in "blended reality": moving interactions out of the screen. In CHI '06 Extended Abstracts on Human Factors in Computing Systems (CHI EA '06). ACM, New York, NY, USA, 893-898. DOI=10.1145/1125451.1125625 Franco Zambonelli and Marco Mamei. 2002. The Cloak of Invisibility: Challenges and Applications. IEEE Pervasive Computing 1, 4 (October 2002), 62-70. DOI=10.1109/MPRV.2002.1158280

3.

Figure 7. Camera position indicated by yellow circle has changed due to blockman’s orientation inside the real lab

4.

5.

Figure 8. View of camera positioned on the yellow circle from figure 7.

6.

Citations
1. Avishek Chatterjee, Suraj Jain, and Venu Madhav Govindu. 2012. A pipeline for building 3D models using depth cameras. In Proceedings of the Eighth Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP '12). ACM, New York, NY, USA, , Article 38 , 8 pages. DOI=10.1145/2425333.2425371

7.