Discovery and Segmentation of Activities in Video

By Matthew Brand, Vera M. Kettnaker. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000

377 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/bdc8cb16e2a0b7a35eec4e891bf6b76871765a5e

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

In Signal Processing

Citations in the corpus

Segmenting Motion Capture Data into Distinct Behaviors (2004, methodology): “HMMs and other clustering techniques bring added power to the problem, but require an expensive search process and/or a very good initial guess.”
Layered representations for learning and inferring office activity from multiple sensory channels (2004, methodology): “Section 5 presents the architecture and implementation of a system named SEER that uses LHMMs, and describes the details of feature extraction, learning and classification used in the system.”
Learning activity patterns using fuzzy self-organizing neural network (2004, background): “The CHMMs are shown to work much more efficiently and accurately than HMMs. Brand et al. [13] show that by minimizing the entropy of the joint distribution, a HMMs internal state machine can be made to organize observed activities into meaningful states.”
Intelligent multi-camera video surveillance: A review (2013, background): “Therefore, people prefer unsupervised approaches (Brand and Kettnaker , 2000; Song et al., 2003; Wang et al., 2006, 2009) which can automatically learn the models of activities without labeling training samples.”
Hidden Markov Models with Kernel Density Estimation of Emission Probabilities and their Use in Activity Recognition (2007, methodology): “In the proposed approach, kernel density estimation of the emission probabilities is operated simultaneously with that of all the other model parameters by an adapted Baum-Welch algorithm.”
Temporal Localization of Actions with Actoms (2013, neutral): “These models require a large amount of training examples in order to model all events that might occur, including non-actions.”
Searching Video for Complex Activities with Finite State Models (2007, methodology): “Methods include: coupled HMM’s ([10]; to classify T’ai Chi moves); layered HMM’s ([22]; to represent office activity); hierachies ([21]; to recognize everyday gesture); HMM’s with a global free parameter ([29]; to model gestures); and entropic HMM’s ([9]; for video puppetry).”
A system for learning statistical motion patterns (2006, background): “Brand and Kettnaker [6] show that, by minimizing the entropy of the joint distribution, observed behaviors can be organized into meaningful states within an HMM. Oliver et al. [34] propose and compare two different state-based learning architectures, namely, HMMs and CHMMs (Coupled Hidden Markov…”
A survey on visual surveillance of object motion and behaviors (2004, methodology): “The CHMMs are shown to work much more efficiently and accurately than HMMs. Brand et al. [134] show that, by the use of the entropy of the joint distribution to learn the HMM, a HMM’s internal state machine can be made to organize observed behaviors into meaningful states.”
Trajectory-Based Anomalous Event Detection (2008, background): “Most of these works rely on the use of hidden Markov models for event recognition, as in the works by Oliver et al. [10], Galata et al. [11], Brand and Kettnaker [12], and Bashir et al. [13].”
Understanding Video Events: A Survey of Methods for Automatic Interpretation of Semantic Occurrences in Video (2009, background): “Another approach to reducing the parameters to be learned is altering the network topology (specifically which states are reachable from which other states) [114].”
Modeling transition patterns between events for temporal human action segmentation and classification (2015, methodology): “Valstar et al. presented a hybrid SVM/HMM system to segment a facial action into temporal phases (e.g. onset, offset, peak, and neutral states), with a noticeable performance gain [23], [24].”
Expandable Data-Driven Graphical Modeling of Human Actions Based on Salient Postures (2008, methodology): “Ahmad and Lee [40] built multiple HMMs for each action, each HMM being for the action observed from a particular viewpoint.”
Understanding Transit Scenes: A Survey on Human Behavior-Recognition Algorithms (2010, methodology): “Bayes network [101], and anomalous behavior detection using trajectory learning with hidden Markov models [102], [103].”
Joint segmentation and classification of human actions in video (2011, methodology): “Brand and Kettnaker [3] trained HMM with entropy minimization and interpreted the hidden states for discovery and segmentation of activities in video.”