The detection of handguns from live-video in real-time based on deep learning

Ghazal, Mohammed; Waisi, Najwan; Abdullah, Nawal Younis

doi:10.12928/telkomnika.v18i6.16174

Cited by 13 publications

(11 citation statements)

References 21 publications

(24 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MobileNetv2, GoogleNet, and MobileNetv3 are the three. According to the results of this method present in this study [12], MobileNetV3 [15] is appropriate for handgun detection since it provides a perfect match between prediction speed and precision. The proposed model had a training accuracy rate of 96%.…”

Section: Related Workmentioning

confidence: 80%

Object Detection using Deep Learning Approach

Dhar¹

2022

IJRASET

View full text Add to dashboard Cite

The most often utilized strategies for current deep learning models to accomplish a multitude of activities on devices are mobile networks and binary neural networks. In this research, we propose a method for identifying an object using the pretrained deep learning model MobileNet for Single Shot Multi-Box Detector (SSD). This technique is utilized for real-time detection as well as webcams to detect the object in a video feed.To construct the module, we use the MobileNet and SSD frameworks to provide a faster and effective deep learning-based object detection approach.Deep learning has evolved into a powerful machine learning technology that incorporates multiple layers of features or representations of data to get cuttingedge results. Deep learning has demonstrated outstanding performance in a variety of fields, including picture classification, segmentation, and object detection. Deep learning approaches have recently made significant progress in fine-grained picture categorization, which tries to differentiate subordinate-level categories.The major goal of our study is to investigate the accuracy of an object identification method called SSD, as well as the significance of a pre-trained deep learning model called MobileNet.To perform this challenge of detecting an item in an image or video, I used OpenCV libraries, Python, and NumPy. This enhances the accuracy of behavior recognition at a processing speed required for real-time detection and daily monitoring requirements indoors and outdoors.

show abstract

Section: Related Workmentioning

confidence: 80%

Object Detection using Deep Learning Approach

Dhar¹

2022

IJRASET

View full text Add to dashboard Cite

show abstract

“…To evaluate the system performance and efficiency, many experiments have been performed (Experiment 1); the ResNet-50 model is loaded with the pre-trained weight of "ImageNet", to use it as a base model. The ImageNet large scale visual recognition challenge (ILSVRC) is a large visual database containing more than 14 million images included with a total of 20,000 categories, which is used for object category classification and detection [26]- [29]. The result from ResNet-50 presents that both train and test performance are less than 50% as shown in Figure 5, this is due to the ImageNet is not specified for face images and contains various image categories.…”

Section: Experimental Work and Resultsmentioning

confidence: 99%

“…Although the black and white face images are robust in face recognition but they are excluded, The VGGFace2 dataset focused on facial and image variation due to color processing as shown in Figure 4. Five age classes have been included in this study {(00-10), (11)(12)(13)(14)(15)(16)(17)(18)(19)(20), (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35), (36-54), (55-90)}.…”

Section: Databasementioning

confidence: 99%

“…Before loading input images, the working environment need some preparation such as installing the required libraries, then start to load the dataset. A preprocessing data, is required at this stage including cleansing the dataset from low quality images which will confuse the training model, A relabeling age group images to five categories {(00-10), (11)(12)(13)(14)(15)(16)(17)(18)(19)(20), (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35), (36-54), (55-90)}, splitting dataset to 80% train, 10% test and 10% validation and finally, set the batch size. The age estimation model is trained with three pre-trained-weight, and two deep learning algorithms (VGG-Face and ResNet50) as it illustrated in the experimental work.…”

Section: Preprocessing Datasetmentioning

confidence: 99%

See 1 more Smart Citation

Pedestrian age estimation based on deep learning

Abdullah

Ghazal²,

Waisi³

2021

IJEECS

Self Cite

View full text Add to dashboard Cite

The large-scale distribution of camera networks in the traffic area resulted in the increasing popularity of video surveillance systems. As pedestrian detection and tracking are the critical monitoring targets in traffic surveillance, many studies focus on pedestrian detection algorithms across cameras. This paper addressed the effect of using the age estimation based on deep convolution neural network (CNN) as a convenience for pedestrian monitoring who is crossing at intersections. Two popular deep convolutional neural networks (DCNNs) pre-trained models have been used in this work, which have recently achieved the best performance in facial features extraction tasks: VGG-Face and ResNet-50. We combined these two models to increase the efficiency of the proposed system. We ran our experiments to evaluate the system based on the VGGFace2 dataset consisting of 3.31 million face images. From the experimental results, we observed a gap in the detection performances between those age groups: children from (00-10) years and elderly with 55 years and more. Moreover, it noted that the proposed pedestrian age estimation model performance is high, also a good result can be obtained by using the model for new purpose.

show abstract

“…Deep learning has achieved great success in a variety of real-world applications and fields. One of the most essential capabilities of deep learning is object detection [1], [2]. The rapid spread of the Corona virus has forced many countries of the world to issue preventive instructions requiring the wearing of face masks.…”

Section: Introductionmentioning

confidence: 99%

Masked face with facial expression recognition based on deep learning

Abdullah

Alkababji

2022

IJEECS

View full text Add to dashboard Cite

Wearing masks contributed to slowing the spread of the coronavirus disease (COVID-19) as the World Health Organization (WHO) recommended wearing face masks especially with the spreading of virus variants like omicron. Although people accept the idea of wearing these masks, it is still unknown the effect of covering parts of the face on social interaction among people in general and children in particular. Moreover, Social isolation affects emotional moods, which causes stress, sadness, and depression. In the current study, we have been exploring the emotional inferences on faces with and without a mask. The system can pick up the universal emotions: fear, disgust, anger, surprise, contempt, sadness, and happiness. The researchers in deep learning are concerned with global pandemic COVID-19 to enhance public health service. The proposed model is developed with a machine learning algorithm through the Haar feature-based cascade classifiers. The built model can detect people's emotions with mask and without a mask with high accuracy.

show abstract

The detection of handguns from live-video in real-time based on deep learning

Cited by 13 publications

References 21 publications

Object Detection using Deep Learning Approach

Object Detection using Deep Learning Approach

Pedestrian age estimation based on deep learning

Masked face with facial expression recognition based on deep learning

Contact Info

Product

Resources

About