论文信息 - Virtual Try-On Using Kinect and HD Camera

Virtual Try-On Using Kinect and HD Camera

We present a virtual try-on system - EON Interactive Mirror - that employs one Kinect sensor and one High-Definition (HD) Camera. We first overview the major technical components for the complete virtual try-on system. We then elaborate on several key challenges such as calibration between the Kinect and HD cameras, and shoulder height estimation for individual subjects. Quality of these steps is the key to achieving seamless try-on experience for users. We also present performance comparison of our system implemented on top of two skeletal tracking SDKs: OpenNI and Kinect for Windows SDK (KWSDK). Lastly, we discuss our experience in deploying the system in retail stores and some potential future improvements.

[1] Kristina Höök,et al. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems , 2012 .

[2] David Kim,et al. HoloDesk: direct 3d interactions with a situated see-through display , 2012, CHI.

[3] Michael Rohs,et al. ShoeSense: a new perspective on gestural interaction and wearable applications , 2012, CHI.

[4] Andrew Wilson,et al. MirageTable: freehand interaction on a projected augmented reality tabletop , 2012, CHI.

[5] Nassir Navab,et al. Estimating human 3D pose from Time-of-Flight images based on geodesic distances and optical flow , 2011, Face and Gesture 2011.

[6] Hans-Peter Seidel,et al. A data-driven approach for real-time full body pose reconstruction from a depth camera , 2011, 2011 International Conference on Computer Vision.

[7] Andrew W. Fitzgibbon,et al. Real-time human pose recognition in parts from single depth images , 2011, CVPR 2011.

[8] Andrew W. Fitzgibbon,et al. KinectFusion: real-time dynamic 3D surface reconstruction and interaction , 2011, SIGGRAPH '11.

[9] Roel Vertegaal,et al. TeleHuman: effects of 3d perspective on gaze and pose estimation with a life-size cylindrical telepresence pod , 2012, CHI.

[10] Mark Pauly,et al. Realtime performance-based facial animation , 2011, ACM Trans. Graph..

[11] Ruigang Yang,et al. Accurate 3D pose estimation from a single depth image , 2011, 2011 International Conference on Computer Vision.