We Are Family: Joint Pose Estimation of Multiple Persons

By Marcin Eichner, Vittorio Ferrari. 2010

124 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Xiaogang Wang
3. Jitendra Malik
4. Bernt Schiele
5. Andrew Zisserman
6. Cordelia Schmid
7. Dacheng Tao
8. Alan L. Yuille

Citations in the corpus

• Towards Accurate Multi-person Pose Estimation in the Wild (2017, methodology): “A similar multi-person PS with an additional explicit occlusion modeling is proposed by Eichner and Ferrari [13].”
• Realtime Multi-person 2D Pose Estimation Using Part Affinity Fields (2016, background): “Eichner et al. [9] extended pictorial structures to take a set of interacting people and depth ordering into account, but still required a person detector to initialize detection hypotheses.”
• DeeperCut: A Deeper, Stronger, and Faster Multi-person Pose Estimation Model (2016, methodology): “[21] use a detector for initialization and reasoning across people, but rely on simple geometric body part relationships and only reason about person-person occlusions.”
• Monocular human pose estimation: A survey of deep learning-based methods (2020, background): “Upper body pose datasets include Bu ff y Stick-men (Ferrari et al., 2008) (frontal-facing view, from indoor TV show), ETHZ PASCAL Stickmen (Eichner et al., 2009) (frontal-facing view, from PASCAL VOC (Everingham et al., 2010)), We Are Family (Eichner and Ferrari, 2010) (Group photo scenario), Video P”
• Heterogeneous Multi-task Learning for Human Pose Estimation with Deep Convolutional Neural Network (2014, background): “We collect training data from several data sets, including Buffy Stickmen [7], ETHZ Stickmen [4], Leed Sport Pose (LSP [13]), Synchronic Activities Stickmen (SA [6]), Frames Labeled In Cinema (FLIC [19]), We Are Family(WAF) [5].”
• Deep Learning-based Human Pose Estimation: A Survey (2020, background): “Buffy Stickmen [239], ETHZ PASCAL Stickmen [240], We Are Family [241], Video Pose 2 [242] and Sync.”
• 2D Human Pose Estimation: New Benchmark and State of the Art Analysis (2014, background): “Existing benchmarks cover aspects of the human pose estimation task such as sport scenes [12, 21], frontal-facing people [8, 3, 17], people interacting with objects [23], pose estimation in group photos [5] and pose estimation of people performing synchronized activities [4].”
• Benchmarking and Error Diagnosis in Multi-instance Pose Estimation (2017, methodology): “However, none of the currently adopted benchmarks for Multi-Person pose estimation [14, 27, 31] carry out an extensive error and performance analysis specific to this framework, and mostly rely on the metrics from the Single-Person case.”
• Recognizing proxemics in personal photos (2012, methodology): “While there has been much work on articulated pose estimation (see the survey in [18]) including methods that jointly estimate poses of multiple people [4], our joint approach is more robust.”
• PoseTrack: A Benchmark for Human Pose Estimation and Tracking (2017, methodology): “The single-frame multi-person pose estimation setting was introduced in [10] along with “We Are Family (WAF)” dataset.”
• Multi-task Recurrent Neural Network for Immediacy Prediction (2015, methodology): “Dataset Annotation Training Test Proxemics [37] pose and interaction 300 289 WeAreFamily [7] pose 350 171 TVHI [23] interaction 599 714 TVHI+ [13] interaction 1654 1566 Immediacy (ours) interaction, pose and posture attributes 7500 2500”
• 3D Human Pose Estimation = 2D Pose Estimation + Matching (2016, background): “This allows our system to take immediate advantage of advances in 2D pose estimation, such as multi-body analysis [7].”
• Human Pose Estimation Using a Joint Pixel-wise and Part-wise Formulation (2013, background): “Others have proposed modelling dependencies and relationships between multiple people [9], which addresses problem (a), and methods for efficiently sampling from pictorial structures [6, 11, 19].”
• Learning effective human pose estimation from inaccurate annotation (2011, methodology): “We base our approach on an extended pictorial structure model (PSM) [11, 17].”
• Whole-Body Human Pose Estimation in the Wild (2020, background): “As shown in Table 1, there are many datasets separately annotated for localizing the keypoints of body [2,3,11,31,62], hand [13,34,51,56,67] or face [4,26,29,33,49,69].”