By,
Fensa Merry Saj
LBSITW
OVERVIEW
 Introduction
 Related Work
 System Setup
 Reconstruction Approach
 Results
 Conclusion
 References
10/5/2...
INTRODUCTION
 Many computer graphics applications require realistic
3D models of human bodies.
 Depth cameras such as Mi...
10/5/2013 4
RELATED WORK
 Scanning devices based on structured light or laser
scan can capture human body with much high quality,
but...
RELATED WORK(contd…)
A most popular one based on light coding is the Microsoft
Kinect Sensor which is at a price of only $...
RELATED WORK(contd…)
Registration with a semi-template.
Rough template, such as the skeleton model of
articulated object, ...
RELATED WORK( contd…)
 The idea of creating a graph of pairwise alignments
between scans.
 First, pairwise rigid alignme...
SYSTEM SETUP
 Two kinects are used to capture the upper and the
lower part of a human body respectively, without
overlapp...
The setup of our system
10/5/2013 10
RECONSTRUCTION APPROACH
 Denote Di={Mi,Ii},i=1,….n as the captured data, n is the
number of captured frames, Mi is the me...
 An accurate template is unavailable.
 We construct an estimated body shape as the template
mesh T1 from the first frame...
 Suppose Mi, i=1…n forming a cycle. fi,j denotes the
registration that can deform mesh Mi to register with
mesh Mj.
 To ...
Pairwise registration:
 For successive frames Mi and Mi+1,corresponding feature
points are obtained by optical flow in th...
Overview of our reconstruction algorithm
10/5/2013 15
10/5/2013 16
Different 3D full human models generated by the system
10/5/2013 17
RESULTS
 Global non-rigid registration gives better result than
global rigid alignment.
 Since the color image and depth...
Realistic virtual try on experience based on the reconstructed
model.(Left)the try on results;(right)the corresponding mes...
Personalized avatar generated by our system.
The motion of the human body is driven by a given skeleton motion
sequence
10...
CONCLUSION
 The proposed method can deal with non-rigid
alignment and complex occlusions.
 The two stage registration al...
REFERENCES
 [1]Jing Tong; Jin Zhou; Ligang Liu; Zhigeng Pan; Hao
Yan, ”Scanning 3D Full Human Bodies Using
Kinects”,Visua...
10/5/2013 23
10/5/2013 24
Upcoming SlideShare
Loading in …5
×

Scanning 3 d full human bodies using kinects

1,320 views

Published on

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,320
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
34
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Scanning 3 d full human bodies using kinects

  1. 1. By, Fensa Merry Saj LBSITW
  2. 2. OVERVIEW  Introduction  Related Work  System Setup  Reconstruction Approach  Results  Conclusion  References 10/5/2013 2
  3. 3. INTRODUCTION  Many computer graphics applications require realistic 3D models of human bodies.  Depth cameras such as Microsoft Kinects are able to capture depth and image data at video rate.  Kinect is compact, low-price and as easy to use as a video camera. 10/5/2013 3
  4. 4. 10/5/2013 4
  5. 5. RELATED WORK  Scanning devices based on structured light or laser scan can capture human body with much high quality, but is very expensive (about $240,000).  Two main approaches employed in depth camera technology are:- based on the time-of-flight principle, measuring time delay between transmissions of a light pulse(about $8,000). based on light coding; projecting a known infrared pattern onto the scene and determining depth based on the pattern’s deformation. 10/5/2013 5
  6. 6. RELATED WORK(contd…) A most popular one based on light coding is the Microsoft Kinect Sensor which is at a price of only $150.  3 Main Types:- Registration without a template.  This method requires high quality scan data & needs small changes in temporal coherence. Registration with a template.  This method needs a relative accurate template & then uses the template to fit each scan. 10/5/2013 6
  7. 7. RELATED WORK(contd…) Registration with a semi-template. Rough template, such as the skeleton model of articulated object, can be utilized.  The first type requires high quality input data & is computationally expensive; the second one needs an accurate template which is hard to fulfil in many applications.  Here, this system uses the third type. 10/5/2013 7
  8. 8. RELATED WORK( contd…)  The idea of creating a graph of pairwise alignments between scans.  First, pairwise rigid alignment is computed in the geometric level.  Global error distribution then operates on an upper level, where errors are measured in terms of the relative rotations and translations of pairwise alignments.  The graph methods can simultaneously minimize the errors of all views rapidly and do not need all scan in memory. 10/5/2013 8
  9. 9. SYSTEM SETUP  Two kinects are used to capture the upper and the lower part of a human body respectively, without overlapping region, from one direction.  A third kinect is used to capture the middle part of the human body from the opposite direction.  The distance between two sets of Kinects is about 2 meters.  A turntable is put in between them. 10/5/2013 9
  10. 10. The setup of our system 10/5/2013 10
  11. 11. RECONSTRUCTION APPROACH  Denote Di={Mi,Ii},i=1,….n as the captured data, n is the number of captured frames, Mi is the merged mesh & Ii is the corresponding image of the i-th frame respectively.  First, a rough template is constructed.  The template is used to deform the geometry of successive frames pair wisely.  Global registration is performed to distribute errors in the deformation space.  Finally, reconstructed model is generated using Poisson reconstruction method. 10/5/2013 11
  12. 12.  An accurate template is unavailable.  We construct an estimated body shape as the template mesh T1 from the first frame.  It is impossible to use this template to register each frame by geometry fitting but it can track the pairwise deformation of successive frames.  T1={v1^k,k=1…K; K is the number of nodes of T1 (typically 50-60). 10/5/2013 12
  13. 13.  Suppose Mi, i=1…n forming a cycle. fi,j denotes the registration that can deform mesh Mi to register with mesh Mj.  To find the pairwise registration f1,2,f2,3,…fn-1,n,fn,1. Deformation Model:  Suppose we have two meshes Mi & Mj, and template mesh at frame i is Ti, then 10/5/2013 13
  14. 14. Pairwise registration:  For successive frames Mi and Mi+1,corresponding feature points are obtained by optical flow in the corresponding images. Projection to the first frame:  n-1 pair wise deformation is required to recover all the relative position of all frames.(refer fig.6)  The desired pairwise deformation f̂1,2,f̂2,3,…f̂n-1,n,f̂n,1 should meet the following conditions:  1.It is cyclic consistent.  2.The original pairwise deformation is relatively correct, so minimize the weighted square error of the new and old deformation. 10/5/2013 14
  15. 15. Overview of our reconstruction algorithm 10/5/2013 15
  16. 16. 10/5/2013 16
  17. 17. Different 3D full human models generated by the system 10/5/2013 17
  18. 18. RESULTS  Global non-rigid registration gives better result than global rigid alignment.  Since the color image and depth information are captured simultaneously and calibrated, the color information of deformed mesh is generated automatically.  Virtual try on.  Personalized avatar-video games,online shopping , human computer interaction etc. 10/5/2013 18
  19. 19. Realistic virtual try on experience based on the reconstructed model.(Left)the try on results;(right)the corresponding meshes. 10/5/2013 19
  20. 20. Personalized avatar generated by our system. The motion of the human body is driven by a given skeleton motion sequence 10/5/2013 20
  21. 21. CONCLUSION  The proposed method can deal with non-rigid alignment and complex occlusions.  The two stage registration algorithm is efficient and of memory efficiency.  The system can generate convincing 3D human bodies at a much low price.  It has good potential for home oriented VR applications. 10/5/2013 21
  22. 22. REFERENCES  [1]Jing Tong; Jin Zhou; Ligang Liu; Zhigeng Pan; Hao Yan, ”Scanning 3D Full Human Bodies Using Kinects”,Visualization and Computer Graphics, IEEE Transactions on, vol.18, no.4, April 2012.  [2]Srivishnu Satyavolu,Gerd Bruder,Pete Willemsen,Frank Stenicke,”Analysis of IR-based virtual reality tracking using multiple Kinects”,2012 IEEE Virtual Reality,2012. 10/5/2013 22
  23. 23. 10/5/2013 23
  24. 24. 10/5/2013 24

×