Suspicious Action Detection in Intelligent Surveillance System Using Action Attribute Modelling

Abstract

Research in the field of image processing and computer vision for recognition of suspicious activity is growing actively. Surveillance systems play a key role in monitoring of sensitive places such as airports, railway stations, shopping complexes, roads, parking areas, roads, banks. For a human it is very difficult to monitor surveillance videos continually, therefore a smart and intelligent system is required that can do real time monitoring of all activities and can categories between usual and some abnormal activities. In this paper many different abnormal activities has been discussed. More focuses is given to violence activity like hitting, slapping, punching etc. For this large human action dataset like UCF101, Kaggel is required. This paper proposes a method to model violence actions using Gaussian Mixture Model with Universal Attribute Model. In this action vector is used to remove redundant attributes and get a low dimensional relevant action vectors.

Human Activity Recognition: Review, Taxonomy and Open Challenges

“…The proposed method achieved 95.88% accuracy on the video dataset. Mudgal et al [68] proposed a smart and intelligent system to monitor normal and abnormal activities such as hitting, slapping, punching, and so on for realtime monitoring of sensitive locations such as airports, banks, and roads. The researcher combined the Gaussian Mixture Model (GMM) with the Universal Attribute Model and performed state-of-the-art feature vectors on the UCF101 human action dataset.…”

“…In the coming years, real-time biometric user recognition of individuals using cell phone plate forms will be prioritized [26]. According to Mudgal et al [68] the proposed method could generate real-time alerts from a live streaming camera.…”

Smart evaluation of videos for unusual-event identification in automated vehicle monitoring systems

“…Issues such as noisy picture inputs, occlusion, vehicle orientations, various number plate kinds, additional images on number plates, nonstandard sizes, low quality of the camera, and so on make number plate identification and recognition a difficult challenge. Existing systems often assume too simple conditions compared to real-world situations, such as only functioning with stationary cameras at a fixed viewing angle and resolution and only with a fixed license plate template (Mudgal et al 2021). As a result, automated activity identification is a need in many video surveillance uses.…”

Design, Simulation and Performance Evaluation of a Risk-Based Border Management System

“…After Stage 1, the risk of all travelers is evaluated and corresponding risk groups, as shown in Figure2, are determined. Travelers who are classified as Unknown (represented by black arrow), which also includes travelers not providing any personal information voluntarily, and Suspected (represented by red arrow) are navigated to Stage 3, while, on the other hand, travelers classified as Trusted (represented by green arrow) and Neutral (represented by yellow arrow) are navigated to Stage 2; • Stage 2: All the neutral and trusted travelers after Stage 1 are navigated to Stage 2.As described in Section 3.1, this stage could have several components like real-time behavioral analytics (RTBA) (e.g.,[58][59][60]), face recognition of consented travelers[61], and web intelligence[62]. After this stage, all the travelers who are classified as "Unknown" and "Suspected" are navigated to Stage 3.…”