Filters








23,458 Hits in 7.4 sec

Recognizing human actions from still images with latent poses

Weilong Yang, Yang Wang, Greg Mori
2010 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition  
We consider the problem of recognizing human actions from still images.  ...  We propose a novel approach that treats the pose of the person in the image as latent variables that will help with recognition.  ...  The primary goal of this work is to recognize actions from still images.  ... 
doi:10.1109/cvpr.2010.5539879 dblp:conf/cvpr/YangWM10 fatcat:jysuanuvrvcjrb36rwvn7klg6a

An expressive deep model for human action parsing from a single image

Zhujin Liang, Xiaolong Wang, Rui Huang, Liang Lin
2014 2014 IEEE International Conference on Multimedia and Expo (ICME)  
This paper aims at one newly raising task in vision and multimedia research: recognizing human actions from still images.  ...  Addressing these problems, we propose to develop an expressive deep model to naturally integrate human layout and surrounding contexts for higher level action understanding from still images.  ...  Fig. 1 . 1 Recognizing human actions from still images, with the help of surrounding objects.  ... 
doi:10.1109/icme.2014.6890158 dblp:conf/icmcs/LiangWHL14 fatcat:k6lvbo2uhzfsnkpmgdep3qdtwy

Latent Body-Pose guided DenseNet for Recognizing Driver's Fine-grained Secondary Activities

Ardhendu Behera, Alexander H Keidel
2018 2018 15th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)  
We also bring together ideas from recent works on human pose detection and transfer learning for visual recognition.  ...  The adapted DenseNet integrates these ideas under one framework, where one stream is focused on the latent body pose and the other stream is on appearance information.  ...  Video is a stack of still frames and there is a variety of work in the field of action recognition from static images.  ... 
doi:10.1109/avss.2018.8639158 dblp:conf/avss/BeheraK18 fatcat:ucljrgonjvbxfa5t3o3vhgo56e

Latent Pose Estimator for Continuous Action Recognition [chapter]

Huazhong Ning, Wei Xu, Yihong Gong, Thomas Huang
2008 Lecture Notes in Computer Science  
In training stage, the human pose is not observed in the action data, and the latent pose estimator is learned under the supervision of the labeled action data, instead of image-to-pose data.  ...  To bridge the gap between the high dimensional observations and the random fields, we propose a novel model that replace the observation layer of a traditional random fields model with a latent pose estimator  ...  We call it latent pose estimator, because it is not explicitly estimated from labeled image-to-pose data. Actually, the human poses are not observed in action training data.  ... 
doi:10.1007/978-3-540-88688-4_31 fatcat:zl253t2rxvbijeietj2w7d4oza

Inferring Unseen Views of People

Chao-Yeh Chen, Kristen Grauman
2014 2014 IEEE Conference on Computer Vision and Pattern Recognition  
Furthermore, we demonstrate their value for recognizing actions in unseen views and estimating viewpoint in novel images.  ...  Given images of people organized by their rough viewpoint, we form a 3D appearance tensor indexed by images (pose examples), viewpoints, and image positions.  ...  Recognizing Actions in Unseen Views Next, we use our inferred views to train a system to recognize actions from a viewpoint it never observed in the training images.  ... 
doi:10.1109/cvpr.2014.258 dblp:conf/cvpr/ChenG14a fatcat:5g4wqre5gfa4bli5umwbuntc6a

Temporal Hockey Action Recognition via Pose and Optical Flows [article]

Zixi Cai, Helmut Neher, Kanav Vats, David Clausi, John Zelek
2018 arXiv   pre-print
Recognizing actions in ice hockey using computer vision poses challenges due to bulky equipment and inadequate image quality.  ...  hockey sticks as an extension of human body pose.  ...  [46] develop action representations based on 2D human poses. Fani et al. [11] use 2D pose from stacked hourglass network to infer action from still images. Iqbal et al.  ... 
arXiv:1812.09533v1 fatcat:yc3vxgo2wvbljfa7wf46i7sd4a

A survey on still image based human action recognition

Guodong Guo, Alice Lai
2014 Pattern Recognition  
Recently still image-based human action recognition has become an active research topic in computer vision and pattern recognition.  ...  It focuses on identifying a person's action or behavior from a single image.  ...  From a sketch of human body poses, it was assumed that there is a great similarity among intra-class poses and the matching of poses can recognize actions.  ... 
doi:10.1016/j.patcog.2014.04.018 fatcat:wja5p7wgf5e2jkfst2w33zv5jy

A Review of Human Activity Recognition Methods

Michalis Vrigkas, Christophoros Nikou, Ioannis A. Kakadiaris
2015 Frontiers in Robotics and AI  
Recognizing human activities from video sequences or still images is a challenging task due to problems, such as background clutter, partial occlusion, changes in scale, viewpoint, lighting, and appearance  ...  In particular, we divide human activity classification methods into two large categories according to whether they use data from different modalities or not.  ...  Also, Yang et al. (2010) combined actions and human poses together, treating poses as latent variables, to infer the action label in still images.  ... 
doi:10.3389/frobt.2015.00028 fatcat:ywzq5ej2gbhatg62sp46t3usgi

Discriminative models for static human-object interactions

Chaitanya Desai, Deva Ramanan, Charless Fowlkes
2010 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Workshops  
We focus on the difficult task of recognizing actions from static images and formulate the problem as a latent structured labeling problem.  ...  We advocate an approach to activity recognition based on modeling contextual interactions between postured human bodies and nearby objects.  ...  Ikizler et al [19] present a similar approach to recognizing actions based on human pose estimation.  ... 
doi:10.1109/cvprw.2010.5543176 dblp:conf/cvpr/DesaiRF10 fatcat:gwmgyhyejngnvmbbsghzebiipu

Hockey Action Recognition via Integrated Stacked Hourglass Network

Mehrnaz Fani, Helmut Neher, David A. Clausi, Alexander Wong, John Zelek
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
The first component is the latent pose estimator, the second transforms latent features to a common frame of reference, and the third performs action recognition.  ...  The hourglass network is employed as the base to generate player pose estimation and layers are added to this network to produce action recognition.  ...  a still image.  ... 
doi:10.1109/cvprw.2017.17 dblp:conf/cvpr/FaniNCWZ17 fatcat:5slfxcc6azf2fnfiqnaek4bbwe

Pose primitive based human action recognition in videos or still images

Christian Thurau, Vaclav Hlavac
2008 2008 IEEE Conference on Computer Vision and Pattern Recognition  
This paper presents a method for recognizing human actions based on pose primitives. In learning mode, the parameters representing poses and activities are estimated from videos.  ...  For recognizing pose primitives, we extend a Histogram of Oriented Gradient (HOG) based descriptor to better cope with articulated poses and cluttered background.  ...  Christian Thurau was supported by a grant from the European Community under the EST Marie Curie Project VISIONTRAIN MRTN-CT-2004-005439.  ... 
doi:10.1109/cvpr.2008.4587721 dblp:conf/cvpr/ThurauH08 fatcat:lkj5p4zgpngezdrj5ozcjeeuju

Learning Composite Latent Structures for 3D Human Action Representation and Recognition

Ping Wei, Hongbin Sun, Nanning Zheng
2019 IEEE transactions on multimedia  
In this paper, we propose that latent states have composite attributes and introduce a novel composite latent structure (CLS) model to represent and recognize 3D human actions with skeleton sequences.  ...  A human action is modeled with a hierarchical graph, which represents the action sequence as sequential atomic actions.  ...  [47] used attributes to represent human action properties and jointly modeled attributes and parts to recognize actions in still images. Liu et al.  ... 
doi:10.1109/tmm.2019.2897902 fatcat:w3huj736ijbnljyzf25q23nnoi

UCF-STAR: A Large Scale Still Image Dataset for Understanding Human Actions

Marjaneh Safaei, Pooyan Balouchian, Hassan Foroosh
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
To benchmark and demonstrate the benefits of UCF-STAR as a large-scale dataset, and to show the role of "latent" motion information in recognizing human actions in still images, we present a novel approach  ...  Action recognition in still images poses a great challenge due to (i) fewer available training data, (ii) absence of temporal information.  ...  TSSTN proves that predicting the "latent" temporal information in still images improves action recognition performance.  ... 
doi:10.1609/aaai.v34i03.5653 fatcat:52c6soxtuze7zeso3ebwxdnxfu

Discriminative Hierarchical Modeling of Spatio-temporally Composable Human Activities

Ivan Lillo, Alvaro Soto, Juan Carlos Niebles
2014 2014 IEEE Conference on Computer Vision and Pattern Recognition  
At the lower level, body poses are encoded in a representative but discriminative pose dictionary. At the intermediate level, encoded poses span a space where simple human actions are composed.  ...  This paper proposes a framework for recognizing complex human activities in videos.  ...  Acknowledgements This work was funded by FONDECYT grant 1120720, from CONICYT, Government of Chile, and LAC-CIR grant RFP1212LAC005. I.L. is supported by a PhD studentship from CONICYT.  ... 
doi:10.1109/cvpr.2014.109 dblp:conf/cvpr/LilloSN14 fatcat:cfzy2qanefd3bndh46fht3pbaa

Deep Video Generation, Prediction and Completion of Human Action Sequences [article]

Haoye Cai, Chunyan Bai, Yu-Wing Tai, Chi-Keung Tang
2017 arXiv   pre-print
In the second stage, a skeleton-to-image network is trained, which is used to generate a human action video given the complete human pose sequence generated in the first stage.  ...  In this paper, we focus on human action videos, and propose a general, two-stage deep framework to generate human action videos with no constraints or arbitrary number of constraints, which uniformly address  ...  Thus, we focus on human action videos in this paper, and divide the task into human pose sequence generation (pose space) followed by image generation (pixel space) from the generated human pose sequences  ... 
arXiv:1711.08682v2 fatcat:5pws3qc5h5dbxodgifkven33wi
« Previous Showing results 1 — 15 out of 23,458 results