BlanketGen - A Synthetic Blanket Occlusion Augmentation Pipeline for Motion Capture Datasets

被引：0

作者：

Carmona, Joao ^{[1
,2
]}

Karacsony, Tamas ^{[1
,2
,3
]}

Silva Cunha, Joao Paulo ^{[1
,2
]}

机构：

[1] INESC TEC, Biomed Engn Res Ctr, Porto, Portugal

[2] Univ Porto, Fac Engn FEUP, Porto, Portugal

[3] Carnegie Mellon Univ, Inst Robot, Pittsburgh, PA 15213 USA

来源：

2023 IEEE 7TH PORTUGUESE MEETING ON BIOENGINEERING, ENBENG | 2023年

关键词：

Human pose estimation; Motion capture; Synthetic occlusions; Cloth simulation; Deep learning; Dataset augmentation;

D O I：

10.1109/ENBENG58165.2023.10175320

中图分类号：

R318 [生物医学工程];

学科分类号：

0831 ;

摘要：

Human motion analysis has seen drastic improvements recently, however, due to the lack of representative datasets, for clinical in-bed scenarios it is still lagging behind. To address this issue, we implemented BlanketGen, a pipeline that augments videos with synthetic blanket occlusions. With this pipeline, we generated an augmented version of the pose estimation dataset 3DPW called BlanketGen3DPW. We then used this new dataset to fine-tune a Deep Learning model to improve its performance in these scenarios with promising results. Code and further information are available at https://gitlab.inesctec.pt/brain-lab/brainlab-public/blanket-gen-releases.

引用

页码：112 / 115

页数：4

共 34 条

[21] Learning 3D joint constraints from vision-based motion capture datasets
Murthy P.
Butt H.T.
Hiremath S.
Khoshhal A.
Stricker D.
IPSJ Transactions on Computer Vision and Applications, 2019, 11 (01)
[22] Robust motion capture system against target occlusion using fast level set method
Iwashita, Yumi
Kurazume, Ryo
Hasegawa, Tsutomu
Hara, Kenji
2006 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA), VOLS 1-10, 2006, : 168 - +
[23] PoseAugment: Generative Human Pose Data Augmentation with Physical Plausibility for IMU-Based Motion Capture
Li, Zhuojun
Yu, Chun
Liang, Chen
Shi, Yuanchun
COMPUTER VISION - ECCV 2024, PT XXXII, 2025, 15090 : 55 - 73
[24] Motion Blur Removal for Uav-Based Wind Turbine Blade Images Using Synthetic Datasets
Peng, Yeping
Tang, Zhen
Zhao, Genping
Cao, Guangzhong
Wu, Chao
REMOTE SENSING, 2022, 14 (01)
[25] Diffusion-Based Data Augmentation for Skin Disease Classification: Impact Across Original Medical Datasets to Fully Synthetic Images
Akrout, Mohamed
Gyepesi, Balint
Hollo, Peter
Poor, Adrienn
Kineso, Blaga
Solis, Stephen
Cirone, Katrina
Kawahara, Jeremy
Slade, Dekker
Abid, Latif
Kovacs, Mate
Fazekas, Istvan
DEEP GENERATIVE MODELS, DGM4MICCAI 2023, 2024, 14533 : 99 - 109
[26] High-quality, cost-effective facial motion capture pipeline with 3D Regression
Moser, Lucio
Williams, Mark
Hendler, Darren
Roble, Doug
SIGGRAPH'18: ACM SIGGRAPH 2018 TALKS, 2018,
[27] A Novel Method of Human Joint Prediction in an Occlusion Scene by Using Low-Cost Motion Capture Technique
Niu, Jianwei
Wang, Xiai
Wang, Dan
Ran, Linghua
SENSORS, 2020, 20 (04)
[28] Synthetic augmentation of cancer cell line multi-omic datasets using unsupervised deep learning (vol 15, 10390, 2024)
Cai, Zhaoxiang
Apolinario, Sofia
Baiao, Ana R.
Pacini, Clare
Sousa, Miguel D.
Vinga, Susana
Reddel, Roger R.
Robinson, Phillip J.
Garnett, Mathew J.
Zhong, Qing
Goncalves, Emanuel
NATURE COMMUNICATIONS, 2025, 16 (01)
[29] Exploitation of motion capture data for improved synthetic micro-Doppler signature generation with adversarial learning
Erol, Baris
Gurbuz, Sevgi Z.
Amin, Moeness G.
BIG DATA II: LEARNING, ANALYTICS, AND APPLICATIONS, 2020, 11395
[30] A Deep Learning Segmentation Pipeline for Cardiac T1 Mapping Using MRI Relaxation-based Synthetic Contrast Augmentation
Bhatt, Nitish
Ramanan, Venkat
Orbach, Ady
Biswas, Labonny
Ng, Matthew
Guo, Fumin
Qi, Xiuling
Guo, Lancia
Jimenez-Juan, Laura
Roifman, Idan
Wright, Graham A.
Ghugre, Nilesh R.
RADIOLOGY-ARTIFICIAL INTELLIGENCE, 2022, 4 (06)

← 1 2 3 4 →