Modeling, Detection and Recognition of Spatio-Temporal Events from Vide

本项目旨在设计和分析用于识别图像序列动态模型的算法，以便从视频中检测、分类和识别时空事件。特别是，我们专注于满足某些统计规律性条件（如二阶平稳性）或某些物理约束（如Lambertianreflection）的图像序列（片段）。虽然这并不包括最一般的视频序列，但通过将每个序列分割成满足假设的部分，将从组合性中获得普遍性。我们的模型的目的是实现动态事件的检测、分类和识别，例如现场视频中烟雾、移动的树叶、火灾、行走的人等。

This project addresses the design and analysis of algorithms for theidentification of dynamical models of image sequences for the purpose ofdetection, classification and recognition of spatio-temporal events fromvideo. In particular, we concentrate on (segments of) image sequences thatsatisfy certain statistical regularity conditions, such as second-orderstationarity, or certain physical constraints, such as Lambertianreflection. While this does not cover the most general video sequences,generality will follow from compositionality, by segmenting each sequenceinto portions that do satisfy the assumptions. The purpose of our models isto enable the detection, classification and recognition of dynamic events,such as the presence of smoke, moving foliage, fire, walking humans etc. inlive video.