Space-time behavior based correlation

By Eli Shechtman, Michal Irani. Computer Vision and Pattern Recognition, 2005

382 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/4811a11937448ea5d40a0af36c974e08ab12c08c

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Parsing Videos of Actions with Segmental Grammars (2014, background): “CFGs: Much prior work has explored CFGs for gesture and event recognition [10, 15, 21, 24].”
Unsupervised Activity Perception in Crowded and Complicated Scenes Using Hierarchical Bayesian Models (2008, background): “Although there are actions/ events modeling approaches [24], [25], [26], [27], [28], which allowed one to detect and separate co-occurring activities, they are usually supervised.”
Learning Motion Categories using both Semantic and Structural Information (2007, methodology): “Various correlation methods such as cross correlation between optical flow descriptors [4] and a consistency measure between ST volumes from their local intensity variations [13] have been proposed.”
Retrieving actions in movies (2007, methodology): “These methods often rely on external localization of people prior to action recognition with exception to [16, 3, 17].”
Human Action Recognition With Video Data: Research and Evaluation Challenges (2014, neutral): “spatiotemporal-features-based volumetric analysis [54], [112], [115], [118], [130], [141].”
Learning human actions via information maximization (2008, methodology): “al.[14] proposed a method which measures the degree of consistency by computing the correlation using the local intensity variance.”
Matching Local Self-Similarities across Images and Videos (2007, methodology): “9 shows results of applying our action detection algorithm for detecting a ballet turn in a video clip of [20].”
Tensor Canonical Correlation Analysis for Action Classification (2007, methodology): “Some recent work has analyzed human actions directly in the space-time volume without explicit motion estimation [1, 4, 8, 7].”
Automatic annotation of human actions in video (2009, methodology): “Whereas most of them evaluate re-sults in simple settings [7, 21, 24], our work is more re-lated to [14] that detects actions in a real movie.”
Machine Recognition of Human Activities: A Survey (2008, background): “Shechtman et al [60] present an approach derived from space-time motion based correlation to match actions with a template.”
Actions as Space-Time Shapes (2007, positive): “We chose to demonstrate our method on the ballet movie example used in [25].”
Human Action Recognition in Videos Using Kinematic Features and Multiple Instance Learning (2010, neutral): “Note that work by [3] and [4] can also be categorized under a volume-based representation.”
Actions as space-time shapes (2005, positive): “We chose to demonstrate our method on the ballet movie example used in [17].”
Recognizing human actions using multiple features (2008, methodology): “The approaches based on volumetric analysis of video include [6, 12, 19] and [11].”
Modeling the Temporal Extent of Actions (2010, neutral): “’s “Space-Time Behavior Based Correlation” techniques [16] [17].”
Action Recognition by Hierarchical Mid-Level Action Elements (2015, methodology): “Moreover, we demonstrate the effectiveness of our model in real-world applications such as action recognition in large-scale untrimmed videos and action parsing.”
Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words (2008, background): “Shechtman and Irani (2005) propose a behavior-based correlation to compute the similarity between space-time volumes which allows to find similar dynamic behaviors and actions.”
Discriminative figure-centric models for joint action localization and recognition (2011, background): “Figure-centric template matching approaches based on shape and motion features are another common strategy [21, 5].”
Exploring Weak Stabilization for Motion Feature Extraction (2013, methodology): “We classify image motion into three types using a stationary world coordinate frame and a moving object coordinate frame.”
Recognition and Segmentation of 3-D Human Action Using HMM and Multi-class AdaBoost (2006, methodology): “While some approaches process the video images directly as spatio-temporal volumes [10] , it is common to first detect and track humans to infer their actions [3] [12].”