Review on Deep Learning Approaches for Anomaly Event Detection in Video Surveillance

Electronics, Год журнала: 2022, Номер 12(1), С. 29 - 29

Опубликована: Дек. 22, 2022

In the last few years, due to continuous advancement of technology, human behavior detection and recognition have become important scientific research in field computer vision (CV). However, one most challenging problems CV is anomaly (AD) because complex environment difficulty extracting a particular feature that correlates with event. As number cameras monitoring given area increases, it will vital systems capable learning from vast amounts available data identify any potential suspicious behavior. Then, introduction deep (DL) has brought new development directions for AD. particular, DL models such as convolution neural networks (CNNs) recurrent (RNNs) achieved excellent performance dealing AD tasks, well other domains like image classification, object detection, speech processing. this review, we aim present comprehensive overview those methods using address problem. Firstly, different classifications anomalies are introduced, then architectures used video discussed analyzed, respectively. The revised contributions been categorized by network type, architecture model, datasets, metrics evaluate these methodologies. Moreover, several applications discussed. Finally, outlined challenges future further field.

Язык: Английский

---

Novel Deep Feature Fusion Framework for Multi-Scenario Violence Detection

Computers, Год журнала: 2023, Номер 12(9), С. 175 - 175

Опубликована: Сен. 5, 2023

Detecting violence in various scenarios is a difficult task that requires high degree of generalisation. This includes fights different environments such as schools, streets, and football stadiums. However, most current research on detection focuses single scenario, limiting its ability to generalise across multiple scenarios. To tackle this issue, paper offers new multi-scenario framework operates two environments: fighting locations rugby has three main steps. Firstly, it uses transfer learning by employing pre-trained models from the ImageNet dataset: Xception, Inception, InceptionResNet. approach enhances generalisation prevents overfitting, these have already learned valuable features large diverse dataset. Secondly, combines extracted through feature fusion, which improves representation performance. Lastly, concatenation step first scenario with second train machine classifier, enabling classifier both highly flexible, can incorporate without requiring training scratch additional The Fusion model, incorporates fusion models, obtained an accuracy 97.66% RLVS dataset 92.89% Hockey Concatenation model accomplished 97.64% 92.41% datasets just classifier. allows for classification violent within Furthermore, not limited be adapted tasks.

Язык: Английский

---

A Comprehensive Survey of Machine Learning Methods for Surveillance Videos Anomaly Detection

IEEE Access, Год журнала: 2023, Номер 11, С. 114680 - 114713

Опубликована: Янв. 1, 2023

Video Surveillance Systems (VSSs) are used in a wide range of applications including public safety and perimeter security. They deployed places such as markets, hospitals, schools, banks, shopping malls, offices, smart cities. VSSs generate massive amount surveillance data, significant research has been published on the use machine learning algorithms to handle data. In this paper, we present an extensive overview thorough analysis cutting-edge methods VSSs. Existing surveys approaches video have some drawbacks, lack in-depth algorithms, omission certain methodologies, insufficient critical evaluation, absence recent algorithms. To fill these gaps, survey provides examination most for anomaly detection. A assessment their strengths, weaknesses, applicability well tailored classifications types different domains provided. Our study also offers insights into future development techniques VSS, positioning itself valuable resource both researchers practitioners field. Finally, share our thoughts what learned how it can help with new developments future.

Язык: Английский

---

Abnormal Behavior Detection in Video Surveillance Using Inception-v3 Transfer Learning Approaches

Iraqi Journal of Computer Communication Control and System Engineering, Год журнала: 2023, Номер unknown, С. 210 - 221

Опубликована: Июнь 29, 2023

The use of video surveillance systems has increased due to security concerns and their relatively low cost. Researchers are working create intelligent Closed Circuit Television (CCTV) cameras that can automatically analyze behavior in real-time detect anomalous behaviors prevent dangerous accidents. Deep Learning (DL) approaches, particularly Convolutional Neural Networks (CNNs), have shown outstanding results analysis anomaly detection. This research paper focused on using Inception-v3 transfer learning approaches improve the accuracy efficiency abnormal detection surveillance. network is used classify keyframes a as normal or by utilizing both pre-training fine-tuning extract features from input data develop new classifier. UCF-Crime dataset train evaluate proposed models. performance models was evaluated accuracy, recall, precision, F1 score. fine-tuned model achieved 88.0%, 89.24%, 85.83%, 87.50% for these measures, respectively. In contrast, pre-trained obtained 86.2%, 86.43%, 84.62%, 85.52%, These demonstrate architecture effectively videos, weights layers further model's performance. Index Terms— Abnormal detection, Video surveillance, learning, Transfer InceptionV3.

Язык: Английский

---

Human Action Behavior Recognition in Still Images with Proposed Frames Selection Using Transfer Learning

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(06), С. 47 - 65

Опубликована: Май 16, 2023

One of the most difficult challenges is recognizing human actions., especially in still images where there isn't much movement. Therefore, Using transfer learning strategy, we suggested a technique for identifying action., which consists training some layers deep techniques while freezing others. Also presented way data split, to choose frames because are working on large dataset such as ucf-101, and this method summarized by discovering features each frame, then clustering elements, choosing percentage cluster test data. We used three techniques. They vgg16, inceptionv3, xception. The proposed models have been implemented UCF-101 Dataset. Depending split methods with dataset, random method, inceptionv3 achieved highest accuracy. In contrast, vgg16 least accuracy, accuracy xception was close that inceptionv3. By comparing size good results: attained an 92.5%, inception v3 98.12%, 95.16%. VGG16 network simple, so less accurate. While xception, more extensive complex, space significant, although prominent only trained blocks top layer

Язык: Английский

---

Detection And Count of Human Bodies In a Crowd Scene Based on Enhancement Features By Using The YOLO v5 Algorithm

Iraqi Journal of Computer Communication Control and System Engineering, Год журнала: 2022, Номер unknown, С. 125 - 134

Опубликована: Июнь 30, 2022

Crowd detection has various applications nowadays. However, detecting humans in crowded circumstances is difficult because the features of different objects conflict, making cross-state impossible. Detectors overlapping zone may therefore overreact. The proposal uses YOLO v5 (You Only Look Once) method to improve crowd recognition and counting. This algorithm entirely accurate detects things real-time. idea relies on edge enhancement pre-processing solve feature regions image performance. CrowdHuman data set used train v5. system counts number detect a crowd. Before training, this model enhanced with several filters. distinguishes person inside by utilizing surrounding box head overall body. Therefore, x-coordinated compared Assume detected heads outnumber bodies. A square will be extracted, but not body square. Also, cropping reduces interference between human beings enhances features. Thus, YOLOv5 can it. improves 2.17 4.1 percent, respectively. Index Terms — Detection, crowed, deep learning, v5, enhancement.

Язык: Английский

---

A Survey of Video Action Recognition Based on Deep Learning

Опубликована: Янв. 1, 2024

Язык: Английский

---

Transfer learning with inception V4 for behavioral anomaly detection in CCTV surveillance

STUDIES IN ENGINEERING AND EXACT SCIENCES, Год журнала: 2024, Номер 5(3), С. e12698 - e12698

Опубликована: Дек. 26, 2024

Heightened security concerns and better affordability have driven the proliferation of video surveillance systems. Developing advanced CCTV cameras capable real-time analysis behavioral patterns to find abnormal is highly anticipated will help avert hazardous incidents. Deep Learning approaches, specifically Convolutional Neural Network (CNN) architecture-based been extensively used for anomaly detection tasks in analytics. The study results from research applying diversified Inception V4 transfer learning methodologies accurately efficiently identify activities This utilized framework classify keyframes that are representative normal or behavior. paper elaborate on techniques pre-training fine-tuning, which employ required attributes input information build a specialized predictor. effectiveness presented models was evaluated through experimental studies UCF-Crime data training testing. Metrics, such as accuracy, recall, precision, F1 scores, were employed evaluation criteria assess performance each model. Fine-Tuned (F-T) model demonstrated metrics 930%, 91.74%, 88.33%, 90.01%, whereas Pre-trained (P-T) showed 88.70%, 88.93%, 87.12%, 88.02%, respectively. These findings suggest Transfer (TL), employing architecture, can effectively distinguish between behaviors. Moreover, adjusting weights particular layers fine-tuning improve performance.

Язык: Английский

---

Estimating Human Running Indoor Based on the Speed of Human Detection by Using OpenPose

Lecture notes in networks and systems, Год журнала: 2023, Номер unknown, С. 749 - 761

Опубликована: Янв. 1, 2023

Язык: Английский

---

Archeological Sites Classification Through Partial Imaging and Convolutional Neural Networks

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(07), С. 26 - 38

Опубликована: Июнь 13, 2023

In this paper, a novel approach for classifying archeological sites using publicly available images through the use of Convolutional Neural Networks (CNNs) is presented. To surmount problem having limited amount data to in training and testing CNNs, our employs technique fine tuning. We conducted an experiment with four popular CNN architectures: VGG-16, VGG-19, ResNet50, InceptionV3. The results show that models achieved impressive accuracy up 98% VGG-16 InceptionV3 97% ResNet50 model, while VGG-19 model produced 95%. study demonstrate effectiveness proposed highlight potential deep learning techniques site classification.

Язык: Английский

---