4-dimensional local spatio-temporal features for human activity recognition

By Hao Zhang, Lynne E. Parker. 2011

155 citations recorded by the source corpus.

Subfields: Biomedical Engineering, 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 Biomedical Engineering

1. Guo-qing Guo
2. Wenbin Chen
3. Yingli Tian
4. Y. Zhu
5. Hao Zhang
6. Jun Liu
7. Fei Han
8. Jinhui Tang

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Dacheng Tao
3. G. Wang
4. Yingli Tian
5. Yi Yang
6. Jun Liu
7. Xiaogang Wang
8. Jitendra Malik

Citations in the corpus

• Space-Time Representation of People Based on 3D Skeletal Data: A Review (2016, methodology): “Among a large number of human representation approaches [24], [25], [26], [27], [28], [29], most of the existing 3D based methods can be broadly grouped into two categories: representations based on local features [30], [31] and skeletonbased representations [32] [33] [34].”
• Learning human activities and object affordances from RGB-D videos (2012, background): “Recently, with the availability of inexpensive RGB-D sensors, some works (Li et al., 2010; Ni et al., 2011; Sung et al., 2011; Zhang and Parker, 2011; Sung et al., 2012) consider detecting human activities from RGB-D videos.”
• Anticipating Human Activities Using Object Affordances for Reactive Robotic Response (2013, methodology): “We will do so by sampling several possible graph structures by augmenting the graph in time, each of which we will call an anticipatory temporal conditional random field (ATCRF).”
• One-Shot Person Re-identification with a Consumer Depth Camera (2014, background): “In particular, several groups have proposed databases of human motions, usually making available skeleton and depth data in conjunction with regular RGB input [33, 18, 30, 25, 32, 20].”
• Depth Induced Feature Representation for 4D Human Activity Recognition (2014, methodology): “For better evaluating depth based activity recognition approaches, several activity datasets are collected by using Kinect devices in very recent years [26][34][41].”
• Qualitative and Quantitative Spatio-temporal Relations in Daily Living Activity Recognition (2014, methodology): “An extension to this method is using a probabilistic approach that combines prior domain knowledge to model each activity as a distribution over the codewords and each video as a distribution over the activities [24].”
• Two-person interaction detection using body-pose features and multiple instance learning (2012, background): “As in [22, 32, 37], combined features with color and depth (e.”
• Human activities recognition using depth images (2013, unknown): “Copyright 2013 ACM 978-1-4503-2404-5/13/10 ...$15.00.”
• Human Daily Action Analysis with Multi-view and Color-Depth Data (2012, background): “[12] consider the actions of color and depth information as 4D hyper cuboids to extract local feature points.”
• Evaluating spatiotemporal interest point features for depth-based action recognition (2014, background): “[26] developed a 4D local spatio-temporal feature which combines both intensity and depth information.”
• Fusing Spatiotemporal Features and Joints for 3D Action Recognition (2013, background): “Zhang et al. [30] in their approach extends the STIPs approach by Dollar el al.”
• Simplex-Based 3D Spatio-temporal Feature Description for Action Recognition (2014, methodology): “• Multi-channel 3D features, typically computed in xyt spatio-temporal space and from multiple information channels, such as RGB and depth channels, including Color-SIFT [7], 4D-LST [30], etc.”
• RGB-D Camera-based Daily Living Activity Recognition (methodology): “Different from traditional RGB channel, recognizing activities using depth images is a new trend in recent research [15] [21] [25] [26] especially after Microsoft released its SDK for Kinect cameras in year 2010 [16].”
• RGB-Depth feature for 3D human activity recognition (2013, methodology): “To better evaluate depth-based activity recognition approaches, several activity datasets are collected by using Kinect devices in very recent years [4, 8, 25].”
• Video-Based Human Behavior Understanding: A Survey (2013, background): “Zhang and Parker [164] employ RGB-D sensors on a mobile robotic platform, aiming at efficient interaction between the robot and humans.”
• Activity Recognition for Natural Human Robot Interaction (2014, background): “Zhang et al.[17] followed a vision based approach and proposed a 4D spatio-temporal feature that combined both intensity and depth information by concatenating depth and intensity gradients within a 4D hyper-cuboid.”
• A Survey on Human Motion Analysis from Depth Data (2013, background): “Each action instance has only 3 samples compared with 33 in [16].”
• Bio-inspired predictive orientation decomposition of skeleton trajectories for real-time human activity prediction (2015, background): “Different from conventional action classification [4], [1], several approaches exist in the literature that focus on activity prediction, i.e., inferring ongoing activities before they are finished.”