Instant First-Person Posture Estimation

Abstract : We propose an instant posture estimation technique, which can operate only by a stereo image with small (6 cm) baseline. It does not require a priori information of the target user, background information, and markers. This system allows cameras to move freely, because it operates only with a small stereo camera unit. Moreover, if the input image is replaced with a movie or real-time video, this system can be used as a real-time motion tracker. With our proposed technique, robots and computers will be able to non-verbally communicate with unspecified people as well as pre-registered people. Moreover, this system can output not only the posture but also the body size and clothes. Therefore, proposed technique can also be used as a calibration procedure for other motion tracking algorithms.
Document type :
Conference papers
Complete list of metadatas

Cited literature [1 references]  Display  Hide  Download

https://hal.inria.fr/hal-01055585
Contributor : Hal Ifip <>
Submitted on : Wednesday, August 13, 2014 - 3:11:23 PM
Last modification on : Wednesday, August 16, 2017 - 5:33:54 PM
Long-term archiving on : Tuesday, April 11, 2017 - 7:40:53 PM

File

icec2010_submission_97.pdf
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Takafumi Serizawa, Yasuyuki Yanagida. Instant First-Person Posture Estimation. 9th International Conference on Entertainment Computing (ICEC), Sep 2010, Seoul, South Korea. pp.508-510, ⟨10.1007/978-3-642-15399-0_75⟩. ⟨hal-01055585⟩

Share

Metrics

Record views

240

Files downloads

133