A Multiple Instance Learning and Relevance Feedback Framework for Retrieving Abnormal Incidents in Surveillance Videos

Abstract

<p class="MsoNormal" style="margin: 0cm 0cm 0pt; mso-layout-grid-align: none;"><span style="font-family: TimesNewRomanPS-BoldMT; font-size: 9pt; mso-bidi-font-weight: bold; mso-bidi-font-family: TimesNewRomanPS-BoldMT;"><span style="font-family: Times New Roman;">This paper incorporates coupled hidden Markov models (CHMM) with relevance feedback (RF) and multiple-instance learning (MIL) for retrieving various abnormal events in surveillance videos. CHMM is suitable for modeling not only the object’s behavior itself but also the interactions between objects. In addition, to address the challenges posed by the “semantic gap” between high level human concepts and the machine-readable low level visual features, we introduce relevance feedback (RF) to bridge the semantic gap by progressively collecting feedback from the user, which allows the machine to discover the semantic meanings of an event by exploring the patterns behind lowlevel features. The adopted multiple-instance learning algorithm enables the proposed framework to provide a user-friendly video retrieval platform with the use of queryby- example (QBE) interface. The experimental results show the effectiveness of the proposed framework in detecting “chasing”, “fighting”, and “robbery” events by demonstrating the increase of retrieval accuracy through iterations and comparing with other methods. By tightly integrating these key components in a learning system, we ease the surveillance video retrieval problem. </span></span><em><span style="font-family: TimesNewRomanPS-BoldItalicMT; font-size: 9pt; mso-bidi-font-weight: bold; mso-bidi-font-family: TimesNewRomanPS-BoldItalicMT;"></span></em></p

End-to-End Deep One-Class Learning for Anomaly Detection in UAV Video Stream

“…The first approach is to carry out the training only on normal data and to consider any type of event outside the training phase as abnormal. Another approach, in contrast to the first, is to use only abnormal events for training [ 14 ]. This approach can be effective in identifying a certain type of abnormal events, but presents a high risk of missing abnormal events that are different from those that have been trained.…”

Multiple Principal-Agent Model of SMEs Pollution Regulation under Centralized Treatment

“…There are two main methods adopted by government to regulate SMEs' pollutant emission, one is installing and improving monitor equipment [3][4][5], the other is making regulatory policies, such as the emissions quota trading [6] and emission tax [7], which are mainly effective for large enterprises. Due to the characteristics of SMEs, which are small scale, numerous and widely scattered, treating pollutants by themselves is not only beyond the ability of the most, but also lead to some problems, such as, scale diseconomies, waste of pollutant treatment resources and the failure of reaching the pollutant treatment standards, and the difficulty of regulation.…”