Hybrid Classifiers for Spatio-temporal Real-time Abnormal Behaviors Detection, Tracking, and Recognition in Massive Hajj Crowds
Authors:
Tarik Alafif,
Anas Hadi,
Manal Allahyani,
Bander Alzahrani,
Areej Alhothali,
Reem Alotaibi,
Ahmed Barnawi
Abstract:
Individual abnormal behaviors vary depending on crowd sizes, contexts, and scenes. Challenges such as partial occlusions, blurring, large-number abnormal behavior, and camera viewing occur in large-scale crowds when detecting, tracking, and recognizing individuals with abnormal behaviors. In this paper, our contribution is twofold. First, we introduce an annotated and labeled large-scale crowd abn…
▽ More
Individual abnormal behaviors vary depending on crowd sizes, contexts, and scenes. Challenges such as partial occlusions, blurring, large-number abnormal behavior, and camera viewing occur in large-scale crowds when detecting, tracking, and recognizing individuals with abnormal behaviors. In this paper, our contribution is twofold. First, we introduce an annotated and labeled large-scale crowd abnormal behaviors Hajj dataset (HAJJv2). Second, we propose two methods of hybrid Convolutional Neural Networks (CNNs) and Random Forests (RFs) to detect and recognize Spatio-temporal abnormal behaviors in small and large-scales crowd videos. In small-scale crowd videos, a ResNet-50 pre-trained CNN model is fine-tuned to verify whether every frame is normal or abnormal in the spatial domain. If anomalous behaviors are observed, a motion-based individuals detection method based on the magnitudes and orientations of Horn-Schunck optical flow is used to locate and track individuals with abnormal behaviors. A Kalman filter is employed in large-scale crowd videos to predict and track the detected individuals in the subsequent frames. Then, means, variances, and standard deviations statistical features are computed and fed to the RF to classify individuals with abnormal behaviors in the temporal domain. In large-scale crowds, we fine-tune the ResNet-50 model using YOLOv2 object detection technique to detect individuals with abnormal behaviors in the spatial domain.
△ Less
Submitted 25 July, 2022;
originally announced July 2022.
Resistance against brute-force attacks on stateless forwarding in information centric networking
Authors:
Bander A. Alzahrani,
Martin J. Reed,
Vassilios G. Vassilakis
Abstract:
Line Speed Publish/Subscribe Inter-networking (LIPSIN) is one of the proposed forwarding mechanisms in Information Centric Networking (ICN). It is a stateless source-routing approach based on Bloom filters. However, it has been shown that LIPSIN is vulnerable to brute-force attacks which may lead to distributed denial-of-service (DDoS) attacks and unsolicited messages. In this work, we propose a n…
▽ More
Line Speed Publish/Subscribe Inter-networking (LIPSIN) is one of the proposed forwarding mechanisms in Information Centric Networking (ICN). It is a stateless source-routing approach based on Bloom filters. However, it has been shown that LIPSIN is vulnerable to brute-force attacks which may lead to distributed denial-of-service (DDoS) attacks and unsolicited messages. In this work, we propose a new forwarding approach that maintains the advantages of Bloom filter based forwarding while allowing forwarding nodes to statelessly verify if packets have been previously authorized, thus preventing attacks on the forwarding mechanism. Analysis of the probability of attack, derived analytically, demonstrates that the technique is highly-resistant to brute-force attacks.
△ Less
Submitted 15 July, 2015;
originally announced July 2015.
