A framework for video event classification by modeling temporal context of multimodal features using HMM.

Highlights: [•] An event classification framework using temporal context of multimodal features. [•] Link low-level features and high-level semantics by mid-level interval features. [•] Explore full temporal relations among mid-level features from multiple modalities. [•] Propose a co-occurrence symbol transformation method for HMM classification. [•] Provide more promising baseball event classification results. [ABSTRACT FROM AUTHOR]