A fully integrated violence detection system using CNN and LSTM

Sharma, Shubham; Sudharsan, Bharath; Naraharisetti, Saamaja; Trehan, Vimarsh; Jayavel, Kayalvizhi

doi:10.11591/ijece.v11i4.pp3374-3380

Cited by 12 publications

(11 citation statements)

References 21 publications

(22 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep learning algorithms have proven effective in a variety of computer vision tasks, such as object classification [9], [17], object detection [12], [18], and action recognition [19], [20], including anomaly detection in video surveillance. As already introduced in the previous section, the approaches that have been proposed to tackle this challenge can be grouped into four categories: reconstruction error, future frame prediction, classifiers, and scoring.…”

Section: Deep Learning-based Methods For Anomaly Detectionmentioning

confidence: 99%

Deep learning-based methods for anomaly detection in video surveillance: a review

et al. 2023

View full text Add to dashboard Cite

Detecting anomalous events in videos is one of the most popular computer vision topics. It is considered a challenging task in video analysis due to its definition, which is subjective or context-dependent. Various approaches have been proposed to address the anomaly detection problems. These approaches vary from hand-crafted to deep learning. Many researchers have gone into determining the best approach for effectively detecting anomalies in video streams while maintaining a low false alarm rate. The results proved that approaches based on deep learning offer very interesting results in this field. In this paper, we review a family of video anomaly detection approaches based on deep learning techniques, which are compared in terms of their algorithms and models. Moreover, we have grouped state-of-the-art methods into different categories based on the approach adopted to differentiate between normal and abnormal events, and the underlying assumptions. Furthermore, we also present publicly available datasets and evaluation metrics used in existing works. Finally, we provide a comparison and discussion on the results of various approaches according to different datasets. This paper can be a good starting point for such researchers to understand this field and review existing work related to this topic.

show abstract

Section: Deep Learning-based Methods For Anomaly Detectionmentioning

confidence: 99%

Deep learning-based methods for anomaly detection in video surveillance: a review

et al. 2023

View full text Add to dashboard Cite

show abstract

“…To get the highest violence detection performance on the UBI-Fights data, different architectures are proposed, based on a main concept of detecting the spatio-temporal features for each video to classify it correctly. The proposed architectures are based on integrating the state-of-art techniques such as Convolutional Block Attention Module (CBAM) of spatial attention and channel attention modules [22], into the traditional ways of catching spatio-temporal features (e.g., CNN+LSTM [6], and ConvLSTM [7,9]); to increase the attention about the most violence features per video. In addition, due to the category imbalance of the UBI-Fights data, the categorical focal loss is used as loss function [23].…”

Section: Model Architecturementioning

confidence: 99%

“…In other words, each frame in the video has a dimension of (height, width, channels), which the overall video has a dimisions of (frames, height, width, channels). So, the justification for DenseNet architecture, to be suitable with the 3D case study; by evolving the DenseNet 2D by a TimeDistribution layer to get the temporal features between video frames [6]. In addition, the other performance enhancing 2D layers type such as MaxPool2D; must be replaced by MaxPool3D.…”

Section: Densenet Convolutional 2dmentioning

confidence: 99%

“…Meanwhile, the most popular work in the violence detection, is done based on surveillance videos, to extract spatio-temporal features, that help in identifying the violence cases clearly [4][5][6][7][8][9][10]. Hereby, this paper proposed different architectures that based on extracting the spatio-temporal features by using different techniques (e.g., 3D Convolutional Neural Network (CNN) [11][12][13], Convolutional Long Short-Term Memory (Conv-LSTM) networks [5,7,14,15], integrating transfer learning with LSTM [16] or Conv-LSTM [7]), and adding a combination of attention modules (I.e., channel attention, and spatial attention) into the architectures to increase the overall classification performance [17].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Violence Detection Enhancement by Involving Convolutional Block Attention Modules Into Various Deep Learning Architectures: Comprehensive Case Study for UBI-Fights Dataset

Abbass

Kang

2023

IEEE Access

View full text Add to dashboard Cite

The violence detection in surveillance videos is a complicated task, due to the requirements of extracting the spatio-temporal features in different videos environment, and various videos prospective cases. Hereby, in this paper, different architectures are proposed to perform this task in high performance, by using the UBI-Fights dataset as a comprehensive case study. The proposed architectures are based on involving the Convolutional Block Attention Modules (CBAM) with other simple layers (e.g., ConvLSTM2D or Conv2D&LSTM). In addition, using the Categorical Focal Loss (CFL) as loss function during architectures training to increase the focus on the most important features. To evaluate the proposed architectures, the performance metrics like are Area Under the Curve (AUC), and Equal Error Rate (EER); are mainly used, to declare the architecture ability of identifying the violence correctly, with low interaction value between classes. The performance results declare the ability of the proposed architectures, to achieve higher results that the state of art techniques. For example, the Conv2D&LSTM based architecture, get AUC value of 0.9493, and EER value of 0.0507; that outperform the most of the other proposed ones, and the state of art performance.

show abstract

“…Al Mamun et al [30] and Ahmed et al [36], deep network based methods are proposed to improve the accuracy of object tracking tasks in varying illumination conditions. End to end deep architectures are proposed for anomaly detection and localization in crowded scenes [17]- [19], [28], and [29]. In these methods feature extraction as well as classification is carried out by deep networks.…”

Section: Introductionmentioning

confidence: 99%

Recognition of crowd abnormal activities using fusion of handcrafted and deep features

Pathade

Khambete

2022

IJEECS

View full text Add to dashboard Cite

Constant vigilance is extremely important at crowded public places where some unusual activities such as sudden dispersion or continuous gathering of people may lead to chaotic and disastrous situations. Automatic recognition of such collective activities of people is indeed an important task to ensure people safety. In this view, we propose a novel approach for automatic recognition of crowd merging and sudden dispersion events. The proposed method detects dispersion and merging using fusion of features extracted by deep networks with a novel set of optical flow based and density based handcrafted features. These proposed features are not affected by occlusion and illumination. These features complement the features extracted by deep network and their fusion improves the performance of event recognition by significant amount. The method is tested profoundly on benchmark public datasets as well as private datasets. Abnormal activity recognition data often suffers from high class imbalance. However, the proposed method could successfully recognize sudden dispersion and merging activities on such very small datasets having class imbalance. This proves the effectiveness and robustness of the proposed features. The proposed method also shows better performance than other state of the art methods based on deep networks.

show abstract

A fully integrated violence detection system using CNN and LSTM

Cited by 12 publications

References 21 publications

Deep learning-based methods for anomaly detection in video surveillance: a review

Deep learning-based methods for anomaly detection in video surveillance: a review

Violence Detection Enhancement by Involving Convolutional Block Attention Modules Into Various Deep Learning Architectures: Comprehensive Case Study for UBI-Fights Dataset

Recognition of crowd abnormal activities using fusion of handcrafted and deep features

Contact Info

Product

Resources

About