The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
본 연구에서는 영상기반 보행분석을 이용하여 체성분 관련 건강지표(체지방, 체수분, 근육의 비율 등)를 추정하는 방법을 제안한다. 이 방법은 기존 체성분 측정기를 이용한 개별 측정보다 더 효율적입니다. 구체적으로 우리는 입력이 보행 에너지 이미지(GEI)이고 출력이 건강 지표로 구성된 CNN(컨볼루션 신경망)을 사용하여 딥 러닝 프레임워크를 설계했습니다. 일반적으로 네트워크 매개 변수를 훈련하려면 방대한 양의 훈련 데이터가 필요하지만, 보행 영상과 피험자별 체성분 측정기를 사용하여 측정한 건강 지표로 구성된 충분한 실측 데이터 쌍을 수집하는 것은 현실적이지 않습니다. 따라서 우리는 많은 수의 주제를 포함하지만 실제 건강 지표가 부족한 보조 보행 데이터 세트를 활용하기 위해 XNUMX단계 접근 방식을 사용합니다. 첫 번째 단계에서는 건강 지표와 관련이 있는 것으로 간주되는 팔 흔들기, 보폭, 구부정한 정도, 몸 너비와 같은 보행 기본 요소를 출력하기 위해 보조 데이터 세트를 사용하여 백본 네트워크를 사전 훈련합니다. 두 번째 단계에서는 백본 네트워크에 일부 레이어를 추가하고 전체 네트워크를 미세 조정하여 상태 표시기의 실측 데이터 포인트 수가 제한되어 있어도 상태 표시기를 출력합니다. 실험 결과는 제안된 방법이 처음부터 훈련할 때뿐만 아니라 자동 인코더 기반 사전 훈련 및 미세 조정 접근 방식을 사용할 때 다른 방법보다 성능이 우수하다는 것을 보여줍니다. 체지방 관련 건강지표를 제외하고 체성분 관련 건강지표에 대해 상대적으로 높은 추정 정확도를 달성하고 있습니다.
Ruochen LIAO
Osaka University
Kousuke MORIWAKI
Osaka University
Yasushi MAKIHARA
Osaka University
Daigo MURAMATSU
Osaka University,Seikei University
Noriko TAKEMURA
Osaka University
Yasushi YAGI
Osaka University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
부
Ruochen LIAO, Kousuke MORIWAKI, Yasushi MAKIHARA, Daigo MURAMATSU, Noriko TAKEMURA, Yasushi YAGI, "Health Indicator Estimation by Video-Based Gait Analysis" in IEICE TRANSACTIONS on Information,
vol. E104-D, no. 10, pp. 1678-1690, October 2021, doi: 10.1587/transinf.2020ZDP7502.
Abstract: In this study, we propose a method to estimate body composition-related health indicators (e.g., ratio of body fat, body water, and muscle, etc.) using video-based gait analysis. This method is more efficient than individual measurement using a conventional body composition meter. Specifically, we designed a deep-learning framework with a convolutional neural network (CNN), where the input is a gait energy image (GEI) and the output consists of the health indicators. Although a vast amount of training data is typically required to train network parameters, it is unfeasible to collect sufficient ground-truth data, i.e., pairs consisting of the gait video and the health indicators measured using a body composition meter for each subject. We therefore use a two-step approach to exploit an auxiliary gait dataset that contains a large number of subjects but lacks the ground-truth health indicators. At the first step, we pre-train a backbone network using the auxiliary dataset to output gait primitives such as arm swing, stride, the degree of stoop, and the body width — considered to be relevant to the health indicators. At the second step, we add some layers to the backbone network and fine-tune the entire network to output the health indicators even with a limited number of ground-truth data points of the health indicators. Experimental results show that the proposed method outperforms the other methods when training from scratch as well as when using an auto-encoder-based pre-training and fine-tuning approach; it achieves relatively high estimation accuracy for the body composition-related health indicators except for body fat-relevant ones.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2020ZDP7502/_p
부
@ARTICLE{e104-d_10_1678,
author={Ruochen LIAO, Kousuke MORIWAKI, Yasushi MAKIHARA, Daigo MURAMATSU, Noriko TAKEMURA, Yasushi YAGI, },
journal={IEICE TRANSACTIONS on Information},
title={Health Indicator Estimation by Video-Based Gait Analysis},
year={2021},
volume={E104-D},
number={10},
pages={1678-1690},
abstract={In this study, we propose a method to estimate body composition-related health indicators (e.g., ratio of body fat, body water, and muscle, etc.) using video-based gait analysis. This method is more efficient than individual measurement using a conventional body composition meter. Specifically, we designed a deep-learning framework with a convolutional neural network (CNN), where the input is a gait energy image (GEI) and the output consists of the health indicators. Although a vast amount of training data is typically required to train network parameters, it is unfeasible to collect sufficient ground-truth data, i.e., pairs consisting of the gait video and the health indicators measured using a body composition meter for each subject. We therefore use a two-step approach to exploit an auxiliary gait dataset that contains a large number of subjects but lacks the ground-truth health indicators. At the first step, we pre-train a backbone network using the auxiliary dataset to output gait primitives such as arm swing, stride, the degree of stoop, and the body width — considered to be relevant to the health indicators. At the second step, we add some layers to the backbone network and fine-tune the entire network to output the health indicators even with a limited number of ground-truth data points of the health indicators. Experimental results show that the proposed method outperforms the other methods when training from scratch as well as when using an auto-encoder-based pre-training and fine-tuning approach; it achieves relatively high estimation accuracy for the body composition-related health indicators except for body fat-relevant ones.},
keywords={},
doi={10.1587/transinf.2020ZDP7502},
ISSN={1745-1361},
month={October},}
부
TY - JOUR
TI - Health Indicator Estimation by Video-Based Gait Analysis
T2 - IEICE TRANSACTIONS on Information
SP - 1678
EP - 1690
AU - Ruochen LIAO
AU - Kousuke MORIWAKI
AU - Yasushi MAKIHARA
AU - Daigo MURAMATSU
AU - Noriko TAKEMURA
AU - Yasushi YAGI
PY - 2021
DO - 10.1587/transinf.2020ZDP7502
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E104-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2021
AB - In this study, we propose a method to estimate body composition-related health indicators (e.g., ratio of body fat, body water, and muscle, etc.) using video-based gait analysis. This method is more efficient than individual measurement using a conventional body composition meter. Specifically, we designed a deep-learning framework with a convolutional neural network (CNN), where the input is a gait energy image (GEI) and the output consists of the health indicators. Although a vast amount of training data is typically required to train network parameters, it is unfeasible to collect sufficient ground-truth data, i.e., pairs consisting of the gait video and the health indicators measured using a body composition meter for each subject. We therefore use a two-step approach to exploit an auxiliary gait dataset that contains a large number of subjects but lacks the ground-truth health indicators. At the first step, we pre-train a backbone network using the auxiliary dataset to output gait primitives such as arm swing, stride, the degree of stoop, and the body width — considered to be relevant to the health indicators. At the second step, we add some layers to the backbone network and fine-tune the entire network to output the health indicators even with a limited number of ground-truth data points of the health indicators. Experimental results show that the proposed method outperforms the other methods when training from scratch as well as when using an auto-encoder-based pre-training and fine-tuning approach; it achieves relatively high estimation accuracy for the body composition-related health indicators except for body fat-relevant ones.
ER -