Authors: Kiran Mudaraddi, Keerthi T, Likitha Yadav, M. Varsha, Nivedita M
DOI Link: https://doi.org/10.22214/ijraset.2023.52381
Certificate: View Certificate
Security is always a main concern in every sphere, due to a rise in crime rate in a crowded event or suspicious lonely areas. Anomaly detection and observance have major applications of computer vision to gear various problems. Due to demand in the protection of safety, security of private properties placement of surveillance systems can recognize and interpret the scene and anomaly events play a vital role in intelligence observance. Detection of weapon and militant using convolution neural network (CNN). Proposed implementation uses two types of datasets. One dataset contains pre-labelled images. And the other one labelled manually contains a set of images. Results are tabulated, both algorithms achieve good efficiency, but their operation in real situations can be based on the trade-off between speed and efficiency. Crime is defined as an act dangerous not only to the person involved, but also to the community as a whole. It is to predict the crime using image dataset and finally calculate accurate performance of the detector. The propose algorithms that are able to alert the human operator when a weapon and militant is visible in the image. It is mainly focused on limiting the number of false alarms in order to allow for real life application of the system. For future work, it is planned to use in live application and to improve the detection and reduce the crime.
I. INTRODUCTION
Currently, numerous cases of crimes are report in public place, home using different types of weapon and militants such as firearms, swords, cutters, etc. To observe and decrease such types of crimes, CCTV camera is installed in public places. Generally, the surveillance footage recorded through these cameras are covered by security staff. Success and failure of detecting crime rely upon the awareness of operator. It is not always possible for a person to pay attention on all the surveillance footage on a single screen recorded through multiple video cameras. Nature and extent of crime depends on the types of weapons and militant that is used.
If a video surveillance system has the ability to generate a prior alert, then by timely reaction losses may be reduced to the maximum extent. Advantage of weapon and militant classification can also be added to a surveillance system. Weapon and militants may be classified either using standard approaches with machine learning classifier or by using deep learning-based approach. The trained model is further used for labelling any new input image. Efficiency of such types of approaches depends on the robustness and diversity of extracted features. To overcome these limitations, deep Convolutional Neural Networks is better to be used as it does not bear any explicit feature of the input image. Deep Convolutional Neural Networks consist of a number of convolutional layers, pooling layers and fully connected layers.
II. LITERATURE SURVEY
III. PROPOSED SYSTEM
Security and protection are a difficult task in today’s modern day world. In order to provide safety for public it is important to have a system that can recognize the unlawful activities.
To tackle this problem we have created a computer based system to identify weapons and militants from the live surveillance camera.
We have implemented this proposed system using YOLOv3, CNN and python to detect weapons and militant. Initially, a dataset is created by collecting images from different resources which consists of various classes of weapons.
This dataset of annotated images of weapons and militant is split training, validation, testing sets and converting into YOLOv3 format.
The YOLOv3 model is then trained on the annotated dataset and evaluated using various metrics such as precision, recall and mAP. Once the model is trained and evaluated, I t is deployed to detect weapons and militants in real time images or videos along with the confidence score of each detected object using python libraries like OpenCV or Pytorch.
IV. METHODOLOGY
In this work, we have attempted to develop an computer based system for security purpose that distinguishes the weapons progressively, if identification is true it will caution the security personals to handle the circumstance by arriving at the place of the incident through surveillance cameras.
Our work starts with collecting the dataset from various sources, then the collected dataset will undergo complete analysis. The image is selected for testing/training purpose only if it matches the requirements and is not repeated. The analysis of image involves pre-processing using YOLOv3 which does image sharpening, labelling of images, removes noise and background subtraction and considers only the image with finer details. Next step is used to extract features from the pre-processed image received as input. In CNN, we take the output from the high-pass filter as input, as CNN is a classifier it has a feature extracting process of its own, using its hidden layers which works in iterations to give a final output.
We have adopted the concept of deep learning that is YOLOv3 and convolutional neural network. YOLOv3 is used to detect the object it takes the entire image at a single time into the CNN and predicts the output through bounding box coordinates and class probabilities. Convolutional neural network is used for feature extraction and classification of the input image.
A. CNN Consists of 4 Layers
a) Convolutional Layer
Convolutional Layer is the first layer in CNN, here 3*3 part of the given matrix which was obtained from High-pass filter is given as input. That 3*3 matrix is multiplied with the High-pass filter matrix for the corresponding position and their sum is written in the particular position.
In this stage one-dimension array is used for the final classification process. The output image obtained from feature extraction is given as input to this process. Where continuous classification of all the features obtained from the previous stage. Each node of the input layer has a value from a one dimension array which represents the feature from the extracted region. That is sent to the hidden layer. Multiple features are getting from the input layer and undergo multiple iteration in the hidden layer. Finally get the predictive values by applying SoftMax activation function to it. Then, get some output values from this process and these values undergo further process. The highest value in the predictive value is considered as output identified as weapon and militant. By using these methods, the weapon and militant will be detected by considering highest accuracy values.
[1] Harsha Jain et.al. “Weapon and militant detection using artificial Intelligence and deep learning for security applications” ICESC 2020. [2] Arif Warsi et.al “Automatic handgun and knife detection algorithms” IEEE Conference 2019. [3] Neelam Dwivedi et.al. “Weapon and militant classification using deep Convolutional neural networks” IEEE Conference CICT 2020. [4] Gyanendra Kumar Verma et.al. “Handheld Gun detection using Faster R-CNN Deep Learning” IEEE Conference 2019. [5] Pallavi Raj et.al. “Simulation and Performance Analysis of Feature Extraction and Matching Algorithms for Image Processing Applications” IEEE International Conference on Intelligent Sustainable Systems, 2019. [6] Mohana et.al. “Simulation of Object Detection Algorithms for Video Surveillance Applications”, International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud), 2018. [7] Deep Learning-based Automatic Weapon Detection System in Video Surveillance\" by Muhammad Tayyab Asif, Muhammad Fraz, and Waseem Hassan. [8] Real-time Weapon Detection Using Deep Learning\" by J. Vanegas and J.Varona in IEEE Access, 2019. [9] \"Automatic Weapon Detection in Surveillance Video Using Convolutional Neural Networks\" by J. He et al. in IEEE Transactions on Multimedia, 2019.
Copyright © 2023 Kiran Mudaraddi, Keerthi T, Likitha Yadav, M. Varsha, Nivedita M. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Paper Id : IJRASET52381
Publish Date : 2023-05-16
ISSN : 2321-9653
Publisher Name : IJRASET
DOI Link : Click Here