Foreground object detection from videos containing complex background
MULTIMEDIA '03 Proceedings of the eleventh ACM international conference on Multimedia
Real-time and accurate segmentation of moving objects in dynamic scene
Proceedings of the ACM 2nd international workshop on Video surveillance & sensor networks
A service-oriented middleware for building context-aware services
Journal of Network and Computer Applications
Nonparametric background generation
Journal of Visual Communication and Image Representation
Fast background subtraction for real time monitoring
ACST'07 Proceedings of the third conference on IASTED International Conference: Advances in Computer Science and Technology
Machine Graphics & Vision International Journal
MODES: moving objects detection and extraction system
ACC'08 Proceedings of the WSEAS International Conference on Applied Computing Conference
Vision Based Pose Recognition in Video Game
Edutainment '08 Proceedings of the 3rd international conference on Technologies for E-Learning and Digital Entertainment
A Local Discriminative Model for Background Subtraction
Proceedings of the 30th DAGM symposium on Pattern Recognition
A hardware architecture for real-time video segmentation utilizing memory reduction techniques
IEEE Transactions on Circuits and Systems for Video Technology
Adaptive background generation for video object segmentation
ISVC'06 Proceedings of the Second international conference on Advances in Visual Computing - Volume Part I
Robust hierarchical multiple hypothesis tracker for multiple-object tracking
Expert Systems with Applications: An International Journal
| Hi-index | 0.00 |
This paper proposes a novel method for detecting foregroundobjects in nonstationary complex environments containingmoving background objects. We derive a Bayesdecision rule for classification of background and foregroundchanges based on inter-frame color co-occurrencestatistics. An approach to store and fast retrieve colorco-occurrence statistics is also established. In the proposedmethod, foreground objects are detected in two steps.First, both foreground and background changes are extractedusing background subtraction and temporal differencing.The frequent background changes are then recognizedusing the Bayes decision rule based on the learnedcolor co-occurrence statistics. Both short-term and longtermstrategies to learn the frequent background changesare proposed. Experiments have shown promising results indetecting foreground objects from video containing waveringtree branches and flickering screens/water surface. Theproposed method has shown better performance as comparedwith two existing methods.