A Study of Body Motion Data Segmentation

碩士 === 國立成功大學 === 機械工程學系 === 102 === With the advancement of motion system, the product, motion data, has widely used in a lot of fields. For the first part, this study introduces the format of storage for motion data. By the method of motion data coding, we can get a BMP file. This files present by...

Full description

Bibliographic Details
Main Authors: Shao-ChiehWen, 溫少捷
Other Authors: Ming-June Tsai
Format: Others
Language:zh-TW
Published: 2014
Online Access:http://ndltd.ncl.edu.tw/handle/91622811676390438964
id ndltd-TW-102NCKU5489125
record_format oai_dc
spelling ndltd-TW-102NCKU54891252016-03-07T04:11:05Z http://ndltd.ncl.edu.tw/handle/91622811676390438964 A Study of Body Motion Data Segmentation 人體運動資料分割研究 Shao-ChiehWen 溫少捷 碩士 國立成功大學 機械工程學系 102 With the advancement of motion system, the product, motion data, has widely used in a lot of fields. For the first part, this study introduces the format of storage for motion data. By the method of motion data coding, we can get a BMP file. This files present by gray or color instead of text. It will provide an intuitive value change and compress archive. For the second part, this study discuss the fairing process. After Mocap experiments, the motion data is needed to do fairing in order to correct the wrong data. In this study, there are four parts for motion data fairing including filtering, interpolation, rectification and smoothing. Filtering and interpolation is used to deal with the missing marker problem. Rectification and smoothing is used to deal with the high-frequency noise. This study will develop some strategies for filtering by the advantages of the model. It can accurately filter off the wrong value. After the fairing process, we can get the smooth motion data and do the motion data segmentation. The concept of segmentation is the analysis of extreme value on motion curve. Extreme value on motion curve is also called the zero-crossing point because the slope is zero. This study will analyze the number of zero-crossing points to develop a method of automatic motion data segmentation. This method can separate a complete motion data into several motion primitives. The motion primitives can be used for motion comparison. Ming-June Tsai 蔡明俊 2014 學位論文 ; thesis 99 zh-TW
collection NDLTD
language zh-TW
format Others
sources NDLTD
description 碩士 === 國立成功大學 === 機械工程學系 === 102 === With the advancement of motion system, the product, motion data, has widely used in a lot of fields. For the first part, this study introduces the format of storage for motion data. By the method of motion data coding, we can get a BMP file. This files present by gray or color instead of text. It will provide an intuitive value change and compress archive. For the second part, this study discuss the fairing process. After Mocap experiments, the motion data is needed to do fairing in order to correct the wrong data. In this study, there are four parts for motion data fairing including filtering, interpolation, rectification and smoothing. Filtering and interpolation is used to deal with the missing marker problem. Rectification and smoothing is used to deal with the high-frequency noise. This study will develop some strategies for filtering by the advantages of the model. It can accurately filter off the wrong value. After the fairing process, we can get the smooth motion data and do the motion data segmentation. The concept of segmentation is the analysis of extreme value on motion curve. Extreme value on motion curve is also called the zero-crossing point because the slope is zero. This study will analyze the number of zero-crossing points to develop a method of automatic motion data segmentation. This method can separate a complete motion data into several motion primitives. The motion primitives can be used for motion comparison.
author2 Ming-June Tsai
author_facet Ming-June Tsai
Shao-ChiehWen
溫少捷
author Shao-ChiehWen
溫少捷
spellingShingle Shao-ChiehWen
溫少捷
A Study of Body Motion Data Segmentation
author_sort Shao-ChiehWen
title A Study of Body Motion Data Segmentation
title_short A Study of Body Motion Data Segmentation
title_full A Study of Body Motion Data Segmentation
title_fullStr A Study of Body Motion Data Segmentation
title_full_unstemmed A Study of Body Motion Data Segmentation
title_sort study of body motion data segmentation
publishDate 2014
url http://ndltd.ncl.edu.tw/handle/91622811676390438964
work_keys_str_mv AT shaochiehwen astudyofbodymotiondatasegmentation
AT wēnshǎojié astudyofbodymotiondatasegmentation
AT shaochiehwen réntǐyùndòngzīliàofēngēyánjiū
AT wēnshǎojié réntǐyùndòngzīliàofēngēyánjiū
AT shaochiehwen studyofbodymotiondatasegmentation
AT wēnshǎojié studyofbodymotiondatasegmentation
_version_ 1718199650051162112