Ontology-Based Analysis of CCTV Data

Abstract

The availability of vast amount of closed-circuit television (CCTV) videos and the critical need for a fast and accurate analysis of all its content especially in security applications, has led to an equally urgent need for its automated processing. Indeed, identifying suspected acts of vandalism or crime from a massive collection of videos are extremely tedious using manual approach. Therefore, with application for security in mind, this paper presents a new approach for automated analysis of CCTV using ontology-based knowledge representation and reasoning. The proposed approach exploits event semantics for the identification of possible crime and perpetrators. The framework consists of salient feature extraction from CCTV footage, inter-level data parsing, highlevel conceptual model event representation, rules inferencing, knowledge reasoning and queries. Although a complete (endto-end) framework was implemented, the main contribution of this work is primarily on the high-level analysis and proposed underpinning Ontology. The proposed approach has been validated with a very large database of CCTV footage from the 2011 London riots. The results validate the strength of the approach.

An aggregated deep convolutional recurrent model for event based surveillance video summarisation: A supervised approach

“…The ontology is built on the famous DOLCE ontology which is extensively used for linguistic and cognitive modelling of knowledge. Kahar and Izquierdo [8] also builds an ontology that aids in identifying major events in CCTV videos. The events for surveillance videos such as crime, attack, smashing, etc.…”

“…The proposed model exhibited better performance by incorporating GRUs over LSTMs. The fundamental equations that govern GRUs are stated in (5)(6)(7)(8).…”

“…The proposed framework identifies five most common classes of events as relevant to surveillance videos. The events that the proposed model is trained to identify are hitting, running, kicking, shooting and punching [7,8]. These actions can be stated as undesirable events which are frequently sought in surveillance videos.…”

“…Efforts to build models to detect the event classes identified in the literature as abnormal or suspicious events such as hit, run, gun shoot, kicking, punching, etc. [7, 8], are minimal. The proposed model is an attempt in this direction to train a model capable of identifying the most common abnormal events in a surveillance scenario.…”

“…Given the recent progress, it is high time a deep learning model that leverages the benefit of supervised learning for abnormal event identification for surveillance video summarisation has been addressed. Efforts to build models to detect the event classes identified in the literature as abnormal or suspicious events such as hit, run, gun shoot, kicking, punching, etc. [7, 8], are minimal. The proposed model is an attempt in this direction to train a model capable of identifying the most common abnormal events in a surveillance scenario. Various datasets have been put forth in the literature for experimenting in the surveillance video summarisation domain.…”