Two-person interaction detection using body-pose features and multiple instance learning

By Kiwon Yun, Jean Honorio, Debaleena Chattopadhyay, Tamara L. Berg, Dimitris Samaras. 2012

540 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Control and Systems Engineering

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

In Control and Systems Engineering

Citations in the corpus

Co-Occurrence Feature Learning for Skeleton Based Action Recognition Using Regularized Deep LSTM Networks (2016, methodology): “We validate the proposed model on the SBU kinect interaction dataset (Yun et al. 2012), HDM05 dataset (M ¨ uller et al. 2007), and CMU dataset (CMU 2003) whose groundtruth was labeled by ourselves.”
RGB-D-based action recognition datasets: A survey (2016, neutral): “Similarly to SBU dataset, K3HI [51]( http://www.lmars.whu.edu.cn:8086/prof_web/zhuxinyan/DataSetPublish/ dataset.html ) is also a two-person interaction dataset.”
Space-Time Representation of People Based on 3D Skeletal Data: A Review (2016, methodology): “The SBU-Kinect-Interaction dataset [132] contains skeleton data of a pair of subjects performing different interaction activities - one person acting and the other reacting.”
Skeleton-Based Online Action Prediction Using Scale Selection Network (2019, neutral): “research domain [26], [27], and a series of hand-crafted features [28], [29], [30], [31], [32], [33], [34], [35], [36], [37] and deep learning-based approaches [18], [19], [38], [39], [40], [41], [42], [43], [44] have been proposed.”
Glimpse Clouds: Human Activity Recognition from Unstructured Feature Points (2018, neutral): “The exclusive usage of pose makes it possible to work on gesture and activity recognition without being a specialist in vision [61, 52], and with significantly reduced dimensionality of the input data.”
RGB-D-based Human Motion Recognition with Deep Learning: A Survey (2017, neutral): “SBU Kinect Interaction Dataset SBU Kinect Interaction Dataset [169](http://www3.”
Skeleton-Based Human Action Recognition With Global Context-Aware Attention LSTM Networks (2017, methodology): “The SBU-Kinect Interaction dataset [72] includes 8 action classes for two-person interaction recognition.”
Skeleton-Based Action Recognition Using Spatio-Temporal LSTM Network with Trust Gates (2017, methodology): “We follow the standard evaluation protocol in [71] and per-”
Modeling Temporal Dynamics and Spatial Configurations of Actions Using Two-Stream Recurrent Neural Networks (2017, methodology): “Following the 5-fold cross validation [53], we split the 21 sets of this dataset into 5 folds and give the average recognition accuracy.”
Pose-conditioned Spatio-Temporal Attention for Human Action Recognition (2017, methodology): “SBU Kinect Interaction Dataset (SBU) [45] —”
Learning and Refining of Privileged Information-Based RNNs for Action Recognition from Depth Sequences (2017, methodology): “We compare the performance of our model with state-ofthe-art methods and baselines on four datasets: MSR Action3D Dataset [26] (Action3D), SBU Interaction dataset [59] (SBU), Cornell Activity Dataset [39](CAD60), and the proposed Blanket dataset (Blanket).”
View Adaptive Recurrent Neural Networks for High Performance Human Action Recognition from Skeleton Data (2017, methodology): “Raw skeleton [46] 49.7 Joint feature [46] 80.3 Raw skeleton [17] 79.4 Joint feature [17] 86.9 HBRNN-L [5] 80.4 Co-occurrence RNN [51] 90.4 STA-LSTM [35] 91.5 ST-LSTM (Tree Traversal) + Trust Gate [24] 93.3 VA-LSTM 97.2 Table 3: Comparisons on the SYSU dataset in accuracy (%).”
PKU-MMD: A Large Scale Benchmark for Continuous Multi-Modal Human Action Understanding (2017, neutral): “G3D [3] 20 210 1467 7 RGB+D+Skeleton Yes 2012 SBU Kinect interaction [54] 8 21 300 14.”
A New Representation of Skeleton Sequences for 3D Action Recognition (2017, methodology): “The proposed method is tested on three skeleton action datasets: NTU RGB+D dataset [37], SBU kinect interaction dataset [53] and CMU dataset [4].”
Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition (2016, methodology): “[59] 80.3% Ji et al., [63] 86.9% CHARM [64] 83.9% HBRNN [30] (reported by [51]) 80.4% Co-occurrence LSTM [51] 90.4% Deep LSTM (reported by [51]) 86.0% ST-LSTM (Joint Chain) 84.7% ST-LSTM (Tree) 88.6% ST-LSTM (Tree) + Trust Gate 93.3% Table 3.”
An End-to-End Spatio-Temporal Attention Model for Human Action Recognition from Skeleton Data (2016, methodology): “We perform our experiments on the following two datasets: the SBU Kinect interaction dataset (Yun et al. 2012), and the largest RGB+D dataset of NTU (Shahroudy et al.”
Online Human Action Detection using Joint Classification-Regression Recurrent Neural Networks (2016, methodology): “Most published RGB-D datasets were generated for the classiﬁcation task where actions were already pre-segmented [27,28].”
Global Context-Aware Attention LSTM Networks for 3D Action Recognition (2017, methodology): “We validate the proposed approach on the NTU RGB+D dataset [35], UT-Kinect dataset [62], and SBU-Kinect Interaction dataset [70].”
Co-occurrence Feature Learning from Skeleton Data for Action Recognition and Detection with Hierarchical Aggregation (2018, methodology): “The SBU Kinect Interaction dataset [ Yun et al. , 2012 ] is a Kinect captured human activity recognition dataset depicting two person interaction.”