A new search and extraction technique for motion capture data
Motion capture is defined as measuring the position and orientation of an object in physical space by triangulating information from multiple cameras which tracks and estimate a number of retro reflective markers over time. This information is later translated into a 3 dimensional digital representa...
Saved in:
| 主要作者: | |
|---|---|
| 格式: | Thesis |
| 語言: | English |
| 出版: |
2008
|
| 主題: | |
| 在線閱讀: | http://eprints.utm.my/id/eprint/9545/1/RafideiMohamadMFSKSM2008.pdf |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 總結: | Motion capture is defined as measuring the position and orientation of an object in physical space by triangulating information from multiple cameras which tracks and estimate a number of retro reflective markers over time. This information is later translated into a 3 dimensional digital representation. Motion capture is applied in a wide range of field such as biomechanics, athletic analysis/training, gait analysis, computer animations, gesture recognition, sign language, music and also fine art dance/performance. In sport science, motion capture data is used for analyzing and perfecting the sequencing mechanics of premiere athletes, as well as monitoring the recovery progress of physical therapies. Existing indexing and extraction technique for motion capture files are based on the whole body motion, where-by motion analysis in sport generally focuses on repeated movements made by specific part of limb such as the arms and legs to measure the effects of a training program. In this research, "Silat Olahraga" movements were used as a case study to apply a new technique for searching and extraction of motion capture files based on different human body segments."Silat Olahraga" is a type of sport based on "Silat", a combative art of Malay fighting and survival. A total of sixteen "Silat Olahraga" motion samples simulating four different motion categories were collected and stored as a motion capture database. The process of identification and extraction of logically related motions scattered within the data set called content-based retrieval method was performed to return results to the user. Results from the experiments show that matching motion files were successfully extracted from the motion capture library using the new algorithm based on different human body segments. |
|---|