Group Activity Detection and Recognition based on sequential Hidden Markov Models
Abstract
Understanding of Group Activities (GA) has significant applications in civilian and military domains. The process of understanding GA is typically involved with spatiotemporal analysis of multi-modality sensor data. Video imagery is one popular sensing modality that offers rich data. However, making sense out of video imagery is a real challenge. In this dissertation a novel Group Activity Detection & Recognition (GADR) System is presented. GADR is a proper inference working model capable of analyzing video imagery frame by frame, extract and inference spatiotemporal information pertaining to observations while developing an incremental perception of the GA as they emerge overtime. A modified Transducer Mark-up Language (TML) data structure is presented to ease the fusion of GA observations and generate semantic messages describing metadata, and attributes of hypothesis under focus. In this report, we propose an ontology based GA recognition where three Hidden Markov Models (HMMs) architectures are used for predicting group activities taking place in outdoor environments and different task operational taxonomy. The three different models include: a concatenated HMM, a cascaded HMM, and a context-based HMM. Experimental results from GA-Hidden Markov Model (GA-HMM) demonstrate the effectiveness and efficiency of this proposed approach for ontology-based group activity detection and recognition. In overall, the proposed GADR system can generate semantic messages with abstract and details describing a monitored GA scenario thereby significantly enhancing Situation Awareness (SA) in Persistent Surveillance Systems.
Subject Area
Computer Engineering|Information Technology|Artificial intelligence
Recommended Citation
Vinayak Elangovan,
"Group Activity Detection and Recognition based on sequential Hidden Markov Models"
(2014).
ETD Collection for Tennessee State University.
Paper AAI3683224.
https://digitalscholarship.tnstate.edu/dissertations/AAI3683224