• Corpus ID: 13117895

An Interval-Based Bayesian Generative Model for Human Complex Activity Recognition

@article{Liu2017AnIB,
  title={An Interval-Based Bayesian Generative Model for Human Complex Activity Recognition},
  author={Li Liu and Yongzhong Yang and Lakshmi Narasimhan Govindarajan and Shu Wang and B. Hu and Li Cheng and David S. Rosenblum},
  journal={ArXiv},
  year={2017},
  volume={abs/1701.00903}
}
Complex activity recognition is challenging due to the inherent uncertainty and diversity of performing a complex activity. Normally, each instance of a complex activity has its own configuration of atomic actions and their temporal dependencies. We propose in this paper an atomic action-based Bayesian model that constructs Allen's interval relation networks to characterize complex activities with structural varieties in a probabilistic generative way: By introducing latent variables from the… 
View PDF on arXiv
View With Semantic Reader

References

SHOWING 1-10 OF 32 REFERENCES
Recognizing Complex Activities by a Probabilistic Interval-Based Model
TLDR
An atomic activity-based probabilistic framework that employs Allen's interval relations to represent local temporal dependencies and introduces a latent variable from the Chinese Restaurant Process to explicitly characterize these unique internal configurations of a particular complex activity as a variable number of tables.
Mining intricate temporal rules for recognizing complex activities of daily living under uncertainty
Modeling Temporal Interactions with Interval Temporal Bayesian Networks for Complex Activity Recognition
TLDR
The interval temporal Bayesian network (ITBN) is introduced, a novel graphical model that combines the Bayesian Network with the interval algebra to explicitly model the temporal dependencies over time intervals and shows significantly improved performance when modeling and recognizing complex activities involving both parallel and sequential events.
Recognizing interleaved and concurrent activities: A statistical-relational approach
TLDR
The use of Markov logic is described as a declarative framework for recognizing interleaved and concurrent activities incorporating both input from pervasive light-weight sensor technology and common-sense background knowledge and its ability to learn statistical-temporal models from training data and to combine these models with background knowledge to improve the overall recognition accuracy.
Discriminative Hierarchical Modeling of Spatio-temporally Composable Human Activities
TLDR
This paper proposes a framework for recognizing complex human activities in videos by formulating model learning in a max-margin framework, and shows how the hierarchical compositional model provides natural handling of occlusions.
Probabilistic event logic for interval-based event recognition
TLDR
This paper argues that holistic reasoning about time intervals of events, and their temporal constraints is critical in such domains to overcome the noise inherent to low-level video representations and proposes a MAP inference algorithm for PEL that addresses the scalability issue of reasoning about an enormous number of time intervals and their constraints in a typical video.
Representation and recognition of action in interactive spaces
TLDR
The PNF propagation algorithm has been applied to an action recognition vision system that handles actions composed of multiple, parallel threads of sub-actions, in situations that can not be efficiently dealt by the commonly used temporal representation schemes such as finite-state machines and HMMs.
Learning Relational Event Models from Video
TLDR
A novel framework (Relational Event Model INDuction) for supervised relational learning of event models from large video datasets using ILP is presented and an extension to the framework is presented by integrating an abduction step that improves the learning performance when there is noise in the input data.
Multiple/Single-View Human Action Recognition via Part-Induced Multitask Structural Learning
TLDR
This paper is the first to demonstrate the applicability of MTSL with part-based regularization on multiple/single-view human action recognition in both RGB and depth modalities.
Semantic Representation and Recognition of Continued and Recursive Human Activities
TLDR
A description-based approach, which enables a user to encode the structure of a high-level human activity as a formal representation, and a system which reliably recognizes sequences of complex human activities with a high recognition rate.
...
1
2
3
4
...