Space-time gestures

By Trevor Darrell, Alex Pentland. 1993

How researchers see this paper

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

1. Pichao Wang
2. Wanqing Li
3. Jun Liu
4. Yonghong Hou
5. Trevor Darrell
6. Hossein Rahmani
7. Dacheng Tao
8. Yi Yang

Citations in the corpus

• Beyond Gaussian Pyramid: Multi-skip Feature Stacking for action recognition (2014, methodology): “For lab datasets where human poses or action templates can be reliably estimated, Dynamic Time Warping (DTP) [5], Hidden Markov Models (HMMs) [41] and Dynamic Bayesian Networks (DBNs) [24] are well studied methods for aligning actions that have speed variation.”
• Action Recognition Based on Joint Trajectory Maps Using Convolutional Neural Networks (2016, background): “ is an important problem in computer vision due to its wide applications in video surveillance, human computer interfaces, robotics, etc. Despite significant research efforts over the past few decades [1]– [21], accurate recognition of human actions from RGB video sequences is still an unsolved probl”
• The Recognition of Human Movement Using Temporal Templates (2001, background): “In contrast to the three-dimensional reconstruction and recognition approaches, others attempt to use only the twodimensional appearance of the action (e.g., [3], [ 10 ], [9], [29]).”
• Recognition of Visual Activities and Interactions by Stochastic Parsing (2000, methodology): “The run-time algorithm used by the HMMs to recognize the words of the gesture vocabulary is a version of [ 16 ] which performs a “backward” match of the signal over a window of a reasonable size.”
• Extraction of 2D Motion Trajectories and Its Application to Hand Gesture Recognition (2002, unknown): “Similarto[10],theyusedthe DynamicTimeWarpingalgorithmtomatchaninputstate sequenceagainstthepreviouslylearnedstatesequencesof gestures.”
• A survey of temporal data mining (2006, background): “The gestures or body motions may represent, e. g., one of a fixed set of messages like waving hello, goodbye, and so on (Darrell & Pentland 1993), or they could be the different strokes in a tennis video (Yamato et al 1992), or in other cases, they could belong to the dictionary of some sign…”
• Task-Specific Gesture Analysis in Real-Time Using Interpolated Views (1996, methodology): “) Using the predictive search pruning mechanism described in [5], which exploits temporal correlation in the observed view scores, we were able to reduce the processing time for this example to under 100ms.”
• Discovering Motion Primitives for Unsupervised Grouping and One-Shot Learning of Human Actions, Gestures, and Expressions (2013, unknown): “Examples of these features include intensity images [6], silhouettes composing motion history (MHI) and motion energy (MEI) images [3], spatiotemporal shapes using tracks of body contours [50], and spatiotemporal volumes spanned by silhouette images [36].”
• Human Action Recognition With Video Data: Research and Evaluation Challenges (2014, background): “Canonical poses [21], person dependent features ([13], [80]) rigid shape formations [36] all suffer from this limitation.”
• Action Recognition Based on Joint Trajectory Maps with Convolutional Neural Networks (2016, background): “Despite significant research efforts over the past few decades [1]– [21], accurate recognition of human actions from RGB video sequences is still an unsolved problem.”
• Recognizing action events from multiple viewpoints (2001, background): “These can be two-dimensional shapes, as in the case of temporal templates[7], view models[ 6 ] or three-dimensional shapes as in the case of XYT cube [14] or image solid[l5].”
• A Review on Human Activity Recognition Using Vision-Based Method (2017, background): “Darrell and Pentland [92] proposed to build the representation of gestures using a set of learned view models.”
• A Unified Framework for Gesture Recognition and Spatiotemporal Gesture Segmentation (2009, background): “In many dynamic gesture recognition systems (e.g., [1], [2], [3], [4], [5], [6], [7]), lower level modules perform spatial and temporal segmentations and extract shape and motion features.”
• A survey on activity recognition and behavior understanding in video surveillance (2013, methodology): “Darrell and Pentland [35] proposed a DTW algorithm and used a view model to recognize gesture actions and effectively handle a variation in the execution of actions.”
• Inferring 3D body pose from silhouettes using activity manifold learning (2004, background): “or may not be tied to specific body parts [ 6 , 5, 31, 13, 27].”
• A review of hand gesture and sign language recognition techniques (2017, methodology): “Darrell and Pentland [52] adapted Dynamic Time Warping (DTW) that had been successfully implemented in speech recognition to recognize dynamic gestures.”