Filters








58 Hits in 4.4 sec

VNect

Dushyant Mehta, Srinath Sridhar, Oleksandr Sotnychenko, Helge Rhodin, Mohammad Shafiei, Hans-Peter Seidel, Weipeng Xu, Dan Casas, Christian Theobalt
2017 ACM Transactions on Graphics  
We present the first real-time method to capture the full global 3D skeletal pose of a human in a stable, temporally consistent manner using a single RGB camera.  ...  Our method's accuracy is quantitatively on par with the best offline 3D monocular RGB pose estimation methods.  ...  We present the first real-time method to capture the full global 3D skeletal pose of a human in a stable, temporally consistent manner using a single RGB camera.  ... 
doi:10.1145/3072959.3073596 fatcat:kviukzjktjbbrfjxcr2kgtytse

XNect: Real-time Multi-person 3D Human Pose Estimation with a Single RGB Camera [article]

Dushyant Mehta, Oleksandr Sotnychenko, Franziska Mueller, Weipeng Xu, Mohamed Elgharib, Pascal Fua, Hans-Peter Seidel, Helge Rhodin, Gerard Pons-Moll, Christian Theobalt
2019 arXiv   pre-print
We present a real-time approach for multi-person 3D motion capture at over 30 fps using a single RGB camera.  ...  The first stage is a convolutional neural network (CNN) that estimates 2D and 3D pose features along with identity assignments for all visible joints of all individuals.  ...  CONCLUSION We present the first real-time approach for multi-person 3D motion capture using a single RGB camera.  ... 
arXiv:1907.00837v1 fatcat:7m5y6d7mtzfnhhgfq56jkm6nl4

A Single RGB Camera Based Gait Analysis with a Mobile Tele-Robot for Healthcare [article]

Ziyang Wang
2020 arXiv   pre-print
As gait analysis with a single camera is much more challenging compared to previous works utilizing multi-cameras, a RGB-D camera or wearable sensors, we propose using vision-based human pose estimation  ...  More specifically, based on the output of two state-of-the-art human pose estimation models (Openpose and VNect), we devise measurements for four bespoke gait parameters: inversion/eversion, dorsiflexion  ...  Finally, the predictions from the 2D heatmap are mapped to 3D global poses. Only a single person can be tracked with VNect in real-time.  ... 
arXiv:2002.04700v4 fatcat:r6rt5u7pkredrpvhhopytxa454

A View-invariant Framework for Fast Skeleton-based Action Recognition using a Single RGB Camera

Enjie Ghorbel, Konstantinos Papadopoulos, Renato Baptista, Himadri Pathak, Girum Demisse, Djamila Aouada, Björn Ottersten
2019 Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications  
The first step is the estimation of a 3D skeleton from a single RGB image using a CNN-based pose estimator such as VNect.  ...  View-invariant action recognition using a single RGB camera represents a very challenging topic due to the lack of 3D information in RGB images.  ...  This CNN pose regression allows the estimation of 2D and 3D skeletons using a monocular RGB camera.  ... 
doi:10.5220/0007524405730582 dblp:conf/visapp/GhorbelPBPDAO19 fatcat:g7565vyywnb6fkw2rrzzx54jf4

View-invariant Action Recognition from Rgb Data via 3D Pose Estimation

Renato Baptista, Enjie Ghorbel, Konstantinos Papadopoulos, Girum G. Demisse, Djamila Aouada, Bjorn Ottersten
2019 ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
In this paper, we propose a novel view-invariant action recognition method using a single monocular RGB camera.  ...  Instead of relying on knowledge transfer, we propose to augment the RGB data by a third dimension by means of 3D skeleton estimation from 2D images using a CNN-based pose estimator.  ...  VNect is based on a CNN pose regression that allows the real-time estimation of 2D and 3D skeletons using a single RGB image.  ... 
doi:10.1109/icassp.2019.8682904 dblp:conf/icassp/BaptistaGPDAO19 fatcat:wyc6yue6vbbxzma4ffh3gztmru

Two-Stage RGB-Based Action Detection Using Augmented 3D Poses [chapter]

Konstantinos Papadopoulos, Enjie Ghorbel, Renato Baptista, Djamila Aouada, Björn Ottersten
2019 Lecture Notes in Computer Science  
In this paper, a novel approach for action detection from RGB sequences is proposed. This concept takes advantage of the recent development of CNNs to estimate 3D human poses from a monocular camera.  ...  Afterwards, to recognize the localized action proposals, a compact Long Short-Term Memory (LSTM) network with a de-noising expansion unit is employed.  ...  VNect has two advantages over alternative 3D pose estimators: real-time performance and temporally-coherent skeletons.  ... 
doi:10.1007/978-3-030-29888-3_3 fatcat:qa23jn2rpbcmndv7e2a7ov3uhy

PhysCap: Physically Plausible Monocular 3D Motion Capture in Real Time [article]

Soshi Shimada, Vladislav Golyanik, Weipeng Xu, Christian Theobalt
2020 arXiv   pre-print
We, therefore, present PhysCap, the first algorithm for physically plausible, real-time and marker-less human 3D motion capture with a single colour camera at 25 fps.  ...  Marker-less 3D human motion capture from a single colour camera has seen significant progress. However, it is a very challenging and severely ill-posed problem.  ...  CONCLUSIONS We have presented PhysCap -the first physics-based approach for a global 3D human motion capture from a single RGB camera that runs in real time at 25 fps.  ... 
arXiv:2008.08880v2 fatcat:vsxoiyd245fcdiyplolq45zrau

Gravity-Aware Monocular 3D Human-Object Reconstruction [article]

Rishabh Dabral and Soshi Shimada and Arjun Jain and Christian Theobalt and Vladislav Golyanik
2021 arXiv   pre-print
., a new approach for joint markerless 3D human motion capture and object trajectory estimation from monocular RGB videos. We focus on scenes with objects partially observed during a free flight.  ...  The proposed human-object interaction constraints ensure geometric consistency of the 3D reconstructions and improved physical plausibility of human poses compared to the unconstrained case.  ...  Estimating Human Poses. We estimate the initial 3D positions of human skeleton joints using the real-time VNect method [24] for the single-person case and XNect [23] for the multi-person setting.  ... 
arXiv:2108.08844v1 fatcat:ve7om7j4kfcfvp76f2vs2gqjqu

Lifting Monocular Events to 3D Human Poses [article]

Gianluca Scarpellini, Pietro Morerio, Alessio Del Bue
2021 arXiv   pre-print
This paper presents a novel 3D human pose estimation approach using a single stream of asynchronous events as input.  ...  Here we propose the first learning-based method for 3D human pose from a single stream of events. Our method consists of two steps.  ...  On the other hand, our approach is the first attempt to estimate 3D human pose based on a single DVS camera. We prove that human pose estimation from event-only DVS camera is feasible.  ... 
arXiv:2104.10609v1 fatcat:tjg6irszhjf47k3ecj56h27vge

Deep3DPose: Realtime Reconstruction of Arbitrarily Posed Human Bodies from Single RGB Images [article]

Liguo Jiang, Miaopeng Li, Jianjie Zhang, Congyi Wang, Juntao Ye, Xinguo Liu, Jinxiang Chai
2021 arXiv   pre-print
We show the system advances the frontier of 3D human body and pose reconstruction from single images by quantitative evaluations and comparisons with state-of-the-art methods.  ...  We introduce an approach that accurately reconstructs 3D human poses and detailed 3D full-body geometric models from single images in realtime.  ...  Our work is also related to model-based tracking of 3D human poses using a single RGB camera.  ... 
arXiv:2106.11536v1 fatcat:2qdgraw3djbxfgidaknnfuv37e

End-to-End Recovery of Human Shape and Pose

Angjoo Kanazawa, Michael J. Black, David W. Jacobs, Jitendra Malik
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
We describe a real time framework for recovering the 3D joint angles and shape of the body from a single RGB image.  ...  Abstract We describe Human Mesh Recovery (HMR), an end-toend framework for reconstructing a full 3D mesh of a human body from a single RGB image.  ...  Tulsiani, A. Kar, S. Gupta, D. Fouhey and Z. Liu for helpful discussions. This research was supported by BAIR sponsors and NSF Award IIS-1526234.  ... 
doi:10.1109/cvpr.2018.00744 dblp:conf/cvpr/KanazawaBJM18 fatcat:awqrxigqizgrheo4y4jwjjuobm

Holistic++ Scene Understanding: Single-view 3D Holistic Scene Parsing and Human Pose Estimation with Human-Object Interaction and Physical Commonsense [article]

Yixin Chen, Siyuan Huang, Tao Yuan, Siyuan Qi, Yixin Zhu, Song-Chun Zhu
2019 arXiv   pre-print
, camera pose, and room layout, and (ii) 3D human pose estimation.  ...  We propose a new 3D holistic++ scene understanding problem, which jointly tackles two tasks from a single-view image: (i) holistic scene parsing and reconstruction---3D estimations of object bounding boxes  ...  3D scene reconstruction and 3D human pose estimation from a single RGB image.  ... 
arXiv:1909.01507v1 fatcat:svbd33j7hvaz5jbysjwwhqhnoy

Holistic++ Scene Understanding: Single-View 3D Holistic Scene Parsing and Human Pose Estimation With Human-Object Interaction and Physical Commonsense

Yixin Chen, Siyuan Huang, Tao Yuan, Yixin Zhu, Siyuan Qi, Song-Chun Zhu
2019 2019 IEEE/CVF International Conference on Computer Vision (ICCV)  
, camera pose, and room layout, and (ii) 3D human pose estimation.  ...  We propose a new 3D holistic ++ scene understanding problem, which jointly tackles two tasks from a single-view image: (i) holistic scene parsing and reconstruction-3D estimations of object bounding boxes  ...  Since VNect can only estimate a single person, we design an additional baseline for 3D multi-person human pose estimation in the world coordinate.  ... 
doi:10.1109/iccv.2019.00874 dblp:conf/iccv/ChenHYZQZ19 fatcat:2dutdksvzjgind7okkv7puymmm

End-to-end Recovery of Human Shape and Pose [article]

Angjoo Kanazawa, Michael J. Black, David W. Jacobs, Jitendra Malik
2018 arXiv   pre-print
We describe Human Mesh Recovery (HMR), an end-to-end framework for reconstructing a full 3D mesh of a human body from a single RGB image.  ...  We do not rely on intermediate 2D keypoint detections and infer 3D pose and shape parameters directly from image pixels. Our model runs in real-time given a bounding box containing the person.  ...  Tulsiani, A. Kar, S. Gupta, D. Fouhey and Z. Liu for helpful discussions. This research was supported by BAIR sponsors and NSF Award IIS-1526234.  ... 
arXiv:1712.06584v2 fatcat:t6bqzrqxq5hevis43fwmj662lq

A Method for Driving Humanoid Robot Based on Human Gesture

Kenta Goto, Hiroaki Nishino, Akihito Yatsuda, Hokuto Tsutsumi, Toshiyuki Haramaki
2020 International Journal of Mechanical Engineering and Robotics Research  
The proposed method captures human gestures in real time by using motion tracking techniques, converting the acquired data to robot motion instructions, and applying them to a physical robot.  ...  It can promote better human robot communication environments with reducing labors in robot motion development.   ...  We use the VNect motion tracking technology for estimating a human pose with a 3D skeleton to drive Pepper, and the OpenPose method for efficiently detecting a pose with a set of 2D feature points to drive  ... 
doi:10.18178/ijmerr.9.3.447-452 fatcat:2lxhcdv2yfe6vmlh3somym3pbi
« Previous Showing results 1 — 15 out of 58 results