Human Motion Capture Berbasis Bebas-Model Menggunakan Penanda Fitur Multi Warna Terparameter

Samuel Gandang Gunanto

Abstract


The utilization of computer vision technology is being developed at this time is the development of science in the realm of digital creative arts, such as animations and games. In this field, computer vision technology plays a role in human motion capture system for generating motion animation of 3D models by real human models through the capture of the camera. With this system of motion generated animation more natural, but the availability of tools and utilization of this technology in the world of animation Indonesia is still very low due to the high price of tools and software used.

The reliability of the system is determined by the accuracy of the estimation of the pose of a model, so that the determination of each segment of the human body in the early stages is the key to success. This research manipulate with features multicolored markers on human motion capture system. Markers are positioned on the joints of human motion in a circle, each color is unique. In addition to cheap and easy to implement, aspects of convenience and flexibility in motion are also taken into consideration the application of these markers.

Calibration Bouguet to 3 cameras used ranged from 0.11 to 0.16 pixels. While the color detection method Giannakopoulos in this case has a value of error of 0.074582. Both values are supporting reconstruction and pose estimation of human stick figures in 3D is quite good, although quantitatively has a point value of the position estimation error of 4.54 cm 3D features.

 

Keywords: multicolor feature, human motion capture, marker based, free models, animation, game.

 

Abstrak 

Pemanfaatan  teknologi  visi  komputer  yang  sedang  berkembang  saat  ini adalah pengembangan di ranah ilmu seni kreatif digital, seperti animasi dan game. Di bidang ini, teknologi visi komputer berperan  pada  sistem penangkapan gerak manusia  untuk  membangkitkan  animasi  gerak  model  3D  oleh model  manusia sesungguhnya  melalui  penangkapan  kamera.  Dengan  sistem  ini  gerak  animasi yang dihasilkan lebih natural, namun ketersediaan alat dan  pemanfaatan teknologi ini di dunia animasi  Indonesia  masih sangat minim dikarenakan mahalnya  harga alat dan perangkat lunak yang dipakai.

Kehandalan  sistem  ini  ditentukan  oleh  ketepatan  estimasi  dari  pose model,  sehingga  penentuan  tiap  segmen  tubuh  manusia  di  tahapan  awal merupakan  kunci  keberhasilannya.  Penelitian  ini  merekayasa  penanda  dengan fitur  multiwarna  pada  sistem  penangkapan  gerak  manusia.  Penanda  diposisikan pada  sendi  gerak  manusia  secara  melingkar  yang  masing-masing  warna  bersifat unik.  Selain  murah  dan  mudah  diterapkan,  aspek  kenyamanan  dan  keluwesan dalam gerak juga menjadi pertimbangan penerapan penanda ini.

Kalibrasi Bouguet untuk 3 kamera yang dipakai berkisar antara 0,11-0,16 piksel.  Sedangkan  deteksi  warna  metode  Giannakopoulos  pada  kasus  ini mempunyai  nilai  kesalahan  sebesar  0,074582. Kedua  nilai  tersebut  mendukung rekonstruksi dan  estimasi pose figur tongkat manusia secara 3D yang cukup baik, meskipun secara kuantitatif memiliki nilai kesalahan estimasi titik posisi fitur 3D sebesar 4,54 cm.

 

Kata Kunci : fitur multiwarna, human motion capture, basis penanda, bebas-model, animasi, game.


Full Text:

ABSTRAK FULL TEXT

References


Aggarwal, J., & Cai., Q. (1999). Human Motion Analysis: A review. Computer Vision and Image Understanding , Vol. 73, no.3, pp. 428-440.

Bouguet, J.-Y. (1999). Visual methods for three dimensional modeling. California: Ph.D Thesis at California Institute of Technology.

Giannakopoulos, T. (2008). Matlab Color Detection Software. Retrieved Mei 12, 2009, from www.di.uoa.gr/~tyiannak

Gunanto, S. G. (2009). Segmentasi Warna Bagian Tubuh Manusia pada Citra 2D. Seminar Nasional Teknologi Informasi dan Aplikasi 2009 (pp. F.133-F.137). Malang: Polinema.

Haritaoglu, I., Harwood, D., & Davis, L. (1998). Ghost: a Human Body Part Labeling System Using Silhouettes. International Conference on Pattern Recognition. Queensland.

Moeslund, T. B. (1999). The Analysis-by-Synthesis Approach in Human Motion Capture: A Review. The 8th Danish conference on pattern recognition and image analysis. Denmark: Copenhagen University.

Moeslund, T. B., Hilton, A., & Kruger, V. (2006). A Survey of Advances in Vision-based Human Motion Capture and Analysis. Computer Vision and Image Understanding , Vol.104, pp. 90-126.

Perales, F. (2002). Human Motion Analysis & Synthesis using Computer Vision and Graphics Techniques:State of Art and Applications. Workshop on Centre of Computer Graphics and Data Visualisation. Czech Republic: University of West Bohemia.

Poppe, R. (2007). Vision-based human motion analysis: An overview. Computer Vision and Image Understanding , Vol. 108, Issues 1-2, pp. 4-18.

Rujikietgumjorn, S. (2008). Segmentation methods for multiple body parts. Knoxville: Project in lieu of Thesis University of Tennessee.

Shipley, T. F., & Brumberg, J. S. (2005). Markerless Motion-capture for Point-light Displays. Philadelphia: Biological Motion Project, Department of Psychology,Temple University.

Starck, J., & Hilton, A. (2003). Model-Based Multiple View Reconstruction of People. Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2.

Tsai, R. Y. (1987). A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf TV cameras and lenses. IEEE Journal of Robotics and Automation RA-3(4) , pp.323-344.

Wachter, S., & Nagel, H. (1999). Tracking Persons in Monocular Image Sequences. Computer Vision and Image Understanding , Vol.74(3), pp. 174-192.

Wren, C. R., Azarbayejani, A., Darrell, T., & P., P. A. (1997). Pfinder: Real-time Tracking of the Human Body. Transactions on Pattern Analysis and Machine Intelligence , Vol.19(7), pp. 780-785.

Zhou, H., & Hu, H. (2004). A Survey-Human Movement Tracking and Stroke Rehabilitation. United Kingdom: Technical Report: Department of Computer Sciences, University of Essex.




DOI: https://doi.org/10.24821/jags.v1i1.898

Article Metrics

Abstract view : 473 times
ABSTRAK - 139 times FULL TEXT - 106 times

Refbacks

  • There are currently no refbacks.




Copyright (c)



Journal of Animation and Games Studies (JAGS) - ISSN 2460-5662 (print) || 2502-499X (online)

Visitor Number  : View JAGS Stats