Video Event Detection Using Motion Relativity and Feature Selection

• Published in: IEEE transactions on multimedia Vol. 16; no. 5; pp. 1303 - 1315
• Main Authors: Wang, Feng, Sun, Zhanhu, Jiang, Yu-Gang, Ngo, Chong-Wah
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.08.2014; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Applied sciences; Artificial intelligence; Cameras; Computer science; control theory; systems; Data processing. List processing. Character string processing; Event detection; Exact sciences and technology; Feature extraction; Feature selection; Gain; Histograms; Invariants; Legged locomotion; Memory organisation. Data processing; motion relativity; Multimedia; Pattern recognition. Digital image processing. Computational geometry; Relativity; Semantics; Software; Tasks; video event detection; Visual; Visualization; Multimedia; Invariant; Computer vision; Dimensionality; Histogram; Feature selection; Event detection; Motion estimation; Image processing; Event management; Image retrieval; Video signal; motion relativity; video event detection; Relative motion; Bag of words; Image analysis; Surveillance; Selection criterion; Motion detection; Content analysis
• ISSN: 1520-9210; 1941-0077
• DOI: 10.1109/TMM.2014.2315780

Event detection plays an essential role in video content analysis. In this paper, we present our approach based on motion relativity and feature selection for video event detection. First, we propose a new motion feature, namely Expanded Relative Motion Histogram of Bag-of-Visual-Words (ERMH-BoW) to employ motion relativity for event detection. In ERMH-BoW, by representing what aspect of an event with Bag-of-Visual-Words (BoW), we construct relative motion histograms between different visual words to depict the objects' activities or how aspect of the event. ERMH-BoW thus integrates both what and how aspects for a complete event description. Meanwhile, we show that by employing motion relativity, ERMH-BoW is invariant to the varying camera movement and able to honestly describe the object activities in an event. Furthermore, compared with other motion features, ERMH-BoW encodes not only the motion of objects, but also the interactions between different objects/scenes. Second, to address the high-dimensionality problem of the ERMH-BoW feature, we further propose an approach based on information gain and informativeness weighting to select a cleaner and more discriminative set of features. Our experiments carried out on several challenging datasets provided by TRECVID for the MED (Multimedia Event Detection) task demonstrate that our proposed approach outperforms the state-of-the-art approaches for video event detection.