Filters








95 Hits in 3.9 sec

Learning 3-D Human Pose Estimation from Catadioptric Videos

Chenchen Liu, Yongzhi Li, Kangqi Ma, Duo Zhang, Peijun Bao, Yadong Mu
2021 Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence   unpublished
3-D human pose estimation is a crucial step for understanding human actions. However, reliably capturing precise 3-D position of human joints is non-trivial and tedious.  ...  In catedioptric videos (\emph{e.g.}, people dance before a mirror), the camera records both the original and mirrored human poses, which provides cues for estimating 3-D positions of human joints.  ...  Acknowledgments This work is supported by National Key R&D Program of China (2018AAA0100702), National Natural Science Foundation of China (61772037) and Beijing Natural Science Foundation (Z190001).  ... 
doi:10.24963/ijcai.2021/118 fatcat:rmycgpcdijeyloddsuvegm7r7a

Constructing panoramic views through facial gaze tracking

Fadi Dornaika, Bogdan Raducanu
2008 2008 IEEE International Conference on Multimedia and Expo  
This paper describes a human machine interaction application for building panoramic views easily and efficiently. The panoramas are not limited to the 1D problem (one axis of rotation).  ...  The resulting system may find applications in online environment mapping as well as in video surveillance. The developed system was applied to map some indoor and outdoor scenes.  ...  Brown and D. Lowe from UBC, Canada [5] .  ... 
doi:10.1109/icme.2008.4607598 dblp:conf/icmcs/DornaikaR08 fatcat:6mft4snc7rf4bloexhpvqszn4a

OmniSCV: An Omnidirectional Synthetic Image Generator for Computer Vision

Bruno Berenguel-Baeta, Jesus Bermudez-Cameo, Jose J. Guerrero
2020 Sensors  
To validate the proposed tool, different computer vision algorithms are tested as line extractions from dioptric and catadioptric central images, 3D Layout recovery and SLAM using equirectangular panoramas  ...  , and 3D reconstruction from non-central panoramas.  ...  (a,b): Real catadioptric images from examples of [34]; (c,d): Synthetic images from our simulator.  ... 
doi:10.3390/s20072066 pmid:32272635 pmcid:PMC7181225 fatcat:te4ofzqlqrcqrg6o4upejkxbfe

3D Human Pose Estimation with a Catadioptric Sensor in Unconstrained Environments Using an Annealed Particle Filter

Fakhreddine Ababsa, Hicham Hadj-Abdelkader, Marouane Boui
2020 Sensors  
Our approach was experimentally evaluated on real data and showed favorable results compared to machine learning based techniques about the 3D pose accuracy.  ...  Thus, the Root Mean Square Error (RMSE) was measured by comparing estimated 3D poses and truth data, resulting in a mean error of 0.065 m when walking action was applied.  ...  In this research work, we propose to estimate, through a particle filter, the 3D human pose from images provided by a catadioptric camera.  ... 
doi:10.3390/s20236985 pmid:33297403 fatcat:dthcnsqjibchxmhhk7sopeugri

MirrorNeRF: One-shot Neural Portrait Radiance Field from Multi-mirror Catadioptric Imaging [article]

Ziyu Wang, Liao Wang, Fuqiang Zhao, Minye Wu, Lan Xu, Jingyi Yu
2021 arXiv   pre-print
We further propose a density regularization scheme to leverage the inherent geometry information from the catadioptric data in a self-supervision manner, which not only improves the training efficiency  ...  camera, which is the first to combine neural radiance field with catadioptric imaging so as to enable one-shot photo-realistic human portrait reconstruction and rendering, in a low-cost and casual capture  ...  Thus d can be computed as in: d = d c − 2(n T d c )n. (5) Finally, we assign the corresponding pixel color to the ray r(t). Fig. 3 illustrates the catadioptric light path in our system.  ... 
arXiv:2104.02607v2 fatcat:okfdrlt255ctrbnyg744m7omvm

Indoor navigation of a non-holonomic mobile robot using a visual memory

Jonathan Courbon, Youcef Mezouar, Philippe Martinet
2008 Autonomous Robots  
During a human-guided learning step, the robot performs paths which are sampled and stored as a set of ordered key images, acquired by an embedded camera.  ...  The proposed framework has been designed for the entire class of central catadioptric cameras (including conventional cameras). It has been validated onto two architectures.  ...  A sequence of images, acquired during a human-guided learning, allows to derive paths driving the robot from its initial to its goal locations.  ... 
doi:10.1007/s10514-008-9093-8 fatcat:pteuv3bx7bc2hmstyhfov6onxm

Structure-From-Motion in Spherical Video Using the von Mises-Fisher Distribution

Hao Guan, William A. P. Smith
2017 IEEE Transactions on Image Processing  
In particular, we propose methods suited to spherical camera geometry for the spherical-n-point problem (estimating camera pose for a spherical image) and calibrated spherical reconstruction (estimating  ...  Index Terms-Structure-from-motion, view stabilisation, spherical image, 360 video.  ...  This is caused by poor estimates of the weights using the noisy pose estimate from the previous iteration.  ... 
doi:10.1109/tip.2016.2621662 pmid:28113757 fatcat:apkamaryefa6rnnw6bk2blcilu

Review on Panoramic Imaging and Its Applications in Scene Understanding [article]

Shaohua Gao, Kailun Yang, Hao Shi, Kaiwei Wang, Jian Bai
2022 arXiv   pre-print
We further offer a detailed analysis of applications of panoramic imaging in scene understanding for autonomous driving and robotics, spanning panoramic semantic image segmentation, panoramic depth estimation  ...  With the rapid development of high-speed communication and artificial intelligence technologies, human perception of real-world scenes is no longer limited to the use of small Field of View (FoV) and low-dimensional  ...  optimization, which achieved more robust pose estimation.  ... 
arXiv:2205.05570v1 fatcat:p77fuqtz2re7zna7kapypcwv5i

Research on Design, Calibration and Real-Time Image Expansion Technology of Unmanned System Variable-Scale Panoramic Vision System

Xiaodong Guo, Zhoubo Wang, Wei Zhou, Zhenhai Zhang
2021 Sensors  
by human eyes.  ...  Compared with the Scaramuzza algorithm, the model improved the calibration accuracy from 5 to 3 pixels.  ... 
doi:10.3390/s21144708 fatcat:5nb6bmo5rjgc3fwt4jglz66stu

Tracking objects with generic calibrated sensors: An algorithm based on color and 3D shape features

M. Taiana, J. Santos, J. Gaspar, J. Nascimento, A. Bernardino, P. Lima
2010 Robotics and Autonomous Systems  
We show tracking applications considering different objects (balls, boxes), several projection models (catadioptric, dioptric, perspective) and several challenging scenarios (clutter, occlusion, illumination  ...  Instead, contour based methods require 3D pose estimation from the 2D silhouette which often involves complex computations and is frequently an ill-posed problem.  ...  Obtained from the projection of the 3D points shown in Fig. 3. (a) For spheres. (b) For cuboids. Fig. 5 .Fig. 6 . 56 Ball tracking in the catadioptric setup.  ... 
doi:10.1016/j.robot.2010.02.010 fatcat:tw3hzrrdqjfntlp26yfuw7i6mq

Camera Models and Fundamental Concepts Used in Geometric Computer Vision

Peter Sturm
2010 Foundations and Trends in Computer Graphics and Vision  
The calibration of the catadioptric sensor then comes down to estimating relative mirror-to-camera pose.  ...  tasks, such as pose and motion estimation.  ... 
doi:10.1561/0600000023 fatcat:i5puai3btrb4vfcvc5zgqgwz3u

Reconstructing 3D Human Pose by Watching Humans in the Mirror [article]

Qi Fang, Qing Shuai, Junting Dong, Hujun Bao, Xiaowei Zhou
2021 arXiv   pre-print
Compared to general scenarios of 3D pose estimation from a single view, the mirror reflection provides an additional view for resolving the depth ambiguity.  ...  In this paper, we introduce the new task of reconstructing 3D human pose from a single image in which we can see the person and the person's image through a mirror.  ...  Related work 2.1. 3D human pose Benefiting from neural networks, the task of monocular 3D human pose estimation has made considerable progress.  ... 
arXiv:2104.00340v1 fatcat:q7bwzclmhrh7zfmneizglol4py

Direct approach to the self-calibration of omnidirectional cameras

A. Salazar-Garibay, E. Malis
2009 2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops  
It is also shown that the uniqueness of the problem can be derived from the solution of non-linear equations.  ...  In this thesis the calibration problem is formalized by using a unified model that is valid for all central catadioptric omnidirectional cameras.  ...  ξ ϕ η = ξ − φ Parabolic 1 1 + 2p −2p Hyperbolic dd 2 +4p 2 < 1 d+2p √ d 2 +4p 2 −2p √ d 2 +4p 2 Elliptical dd 2 +4p 2 < 1 d−2p √ d 2 +4p 2 2p √ d 2 +4p 2 Planar 0 1 -1 d: distance between focal points  ... 
doi:10.1109/iccvw.2009.5457546 dblp:conf/iccvw/Salazar-Garibay09 fatcat:hdluhhwivvbxxkq2pmwpot3h6a

Wide-angle Visual Feature Matching for Outdoor Localization

Peter Hansen, Peter Corke, Wageeh Boles
2009 The international journal of robotics research  
We present results from two monocular wide-angle outdoor image sequences obtained using fisheye and equiangular catadioptric image sequences.  ...  We also demonstrate the use of the technique for variable baseline visual odometry and for place recognition using algorithms inspired by "Video Google".  ...  Importantly, this metric is learned during offline training. Establishing the visual-words Unlike Video Google we do not track features across frames to determine stable features.  ... 
doi:10.1177/0278364909356484 fatcat:7yyibgemcvaxjijcv57ryatitq

Multiperspective Modeling, Rendering, and Imaging [article]

Jingyi Yu, Leonard McMillan, Peter Sturm
2008 Eurographics State of the Art Reports  
They will learn about multiperspective modeling [...]  ...  In contrast, a multiperspective image combines what is seen from several viewpoints into a single image.  ...  For perspective cameras, pose can be estimated, up to a finite number of solutions, from 3 point correspondences (3D-2D) already. The same holds for general cameras.  ... 
doi:10.2312/egst.20081047 fatcat:pjzf222crzcdljozemdehlp3se
« Previous Showing results 1 — 15 out of 95 results