Automated Teller Machines also known as ATM's are widely used nowadays by each and everyone. The ATM machine (Automated Teller Machine) is an electronic device that is used by the banks to perform banking tasks li...
详细信息
Nowadays we are facing a pandemic, there is a situation where people are not ready to wear face masks, or they do not wear them properly, so, in this research, we are introducing an automatic mask detection system usi...
详细信息
Color model is also referred to as a mathematical organization or arrangement of colors as numerical values as three or four color components or channels. Every color space has unique features and application oriented...
详细信息
Independent component analysis (ICA) is an unsupervised learning approach for computing the independent components (ICs) from the multivariate signals or data matrix. The ICs are evaluated based on the multiplication ...
详细信息
Independent component analysis (ICA) is an unsupervised learning approach for computing the independent components (ICs) from the multivariate signals or data matrix. The ICs are evaluated based on the multiplication of the weight matrix with the multivariate data matrix. This study proposes a novel Pt/Cu:ZnO/Nb:STO memristor crossbar array for the implementation of both ACY ICA and Fast ICA for blind source separation. The data input was applied in the form of pulse width modulated voltages to the crossbar array and the weight of the implemented neural network is stored in the memristor. The output charges from the memristor columns are used to calculate the weight update, which is executed through the voltages kept higher than the memristor Set/Reset voltages (+/- 1.30 v). In order to demonstrate its potential application, the proposed memristor crossbar arrays based fast ICA architecture is employed for image source separation problem. The experimental results demonstrate that the proposed approach is very effective to separate image sources, and also the contrast of the images are improved with an improvement factor in terms of percentage of structural similarity as 67.27% when compared with the software-based implementation of conventional ACY ICA and Fast ICA algorithms.
Wildfires pose a significant threat to ecosystems, biodiversity, and human settlements, with climate change and deforestation exacerbating the frequency and intensity. Conventional forest fire detection methods, inclu...
详细信息
ISBN:
(数字)9798331501488
ISBN:
(纸本)9798331501495
Wildfires pose a significant threat to ecosystems, biodiversity, and human settlements, with climate change and deforestation exacerbating the frequency and intensity. Conventional forest fire detection methods, including satellite-based monitoring and ground surveillance, suffer from limitations such as delayed detection, high false alarm rates, and dependency on manual intervention. Existing sensor-based detection systems often lack integration with intelligent decision-making frameworks, leading to inefficient fire prevention strategies. The absence of real-time processing and predictive capabilities further reduces the effectiveness of current methodologies in mitigating wildfire risks. This work presents an IoT-based deep learning approach utilizing Convolutional Neural Networks (CNNs) for real-time forest fire detection and prevention. Data is collected from satellite images, drone feeds, and IoT sensors, including temperature, humidity, gas concentration, and infrared readings. Advanced data preprocessing techniques, including image augmentation, normalization, and feature extraction, enhance the robustness of the model. A CNN-based classification model is implemented to analyze fire patterns and assess risk levels. The system is trained using sensor data and image datasets, ensuring high detection accuracy with minimal false positives. The key features of this methodology include real-time wildfire detection, automated risk assessment, and deployment on IoT edge devices for immediate decision-making. The integration of sensor-driven insights with deep learning improves early detection capabilities while reducing reliance on manual surveillance. Efficient data handling and cloud-based alert mechanisms enable proactive measures against fire outbreaks. The proposed system enhances wildfire monitoring by ensuring fast response times, improved accuracy, and scalable deployment in fire-prone regions.
In semantic segmentation, category-hidden attack is a malicious adversarial attack which manipulates a specific category without affecting the recognition of other objects. A popular method is the nearest-neighbor alg...
详细信息
In semantic segmentation, category-hidden attack is a malicious adversarial attack which manipulates a specific category without affecting the recognition of other objects. A popular method is the nearest-neighbor algorithm, which modifies the segmentation map by replacing a target category with other categories close to it. Nearest-neighbor method aims to restrict the strength of perturbation noise that is imperceptive to both human eyes and segmentation algorithms. However, its spatial search adds lots of computational burden. In this paper, we propose two fast methods, dot-based method and line-based method, which are able to quickly complete the category transfers in logits maps without spatial search. The advantages of our two methods result from generating the logits maps by modifying the probability distribution of the category channels. Both of our methods are global, and the location and size of objects to hide are not cared, so their processing speed is very fast. The dot-based algorithm takes the pixel as the unit of calculation, and the line-based algorithm combines the category distribution characteristics of the horizontal direction to calculate. Experiments verify the effectiveness and efficiency compared with nearest-neighbor method. Specifically, in the segmentation map modification step, our methods are 5 times and 65 times faster than nearest-neighbor, respectively. In the small perturbation attack experiment, dot-based method gets the fastest speed, while different datasets and different setting experiments indicate that the line-based method is able to achieve faster and better adversarial segmentation results in most cases. (c) 2021 The Authors. Published by Atlantis Press B.v. This is an open access article distributed under the CC BY-NC 4.0 license (http://***/licenses/by-nc/4.0/).
The combination of car model classification and accelerated advances in computer vision appears to have the potential to significantly revolutionize automated transportation systems. Pattern recognition and image proc...
The combination of car model classification and accelerated advances in computer vision appears to have the potential to significantly revolutionize automated transportation systems. Pattern recognition and imageprocessing-based car classification methods have been proposed in recent years to improve the performance of smart highway toll collection and monitoring of traffic systems. Unfortunately, all these algorithms are trained on a inadequate number of features derived from tiny datasets, which do not considered for real-time road traffic circumstances. To address these difficulties, proposed system presents a classification model for vehicle based on convolutional neural networks to enhance the resilience of car classifier model in real-time applications. To achieve real-time robustness in classification systems, this proposed system offers a Stanford car dataset of 16,185 images classified into 157 common vehicle classes. To begin, the car dataset was trained using Inception-v3, vGG16, and ResNet50 to evaluate their accuracy based on performance. As a result, a comparative study was carried out to assess the effectiveness of the proposed system.
Computer vision systems and algorithms are designed to process digital images and extract necessary information from it. In this research we propose computer vision system consisting of several spherical mobile device...
详细信息
ISBN:
(纸本)9781728160726
Computer vision systems and algorithms are designed to process digital images and extract necessary information from it. In this research we propose computer vision system consisting of several spherical mobile devices with digital camera and microcomputer inside. Device's design, basic characteristics and current results, as well as steps for further actions to improve the system are described below in the paper. At the core of imageprocessing OpenCv library and modern convolutional neural networks such as YOLOv3, MobileNETSSDv2 are used.
Object detection has seen many changes in algorithms to improve performance both on speed and accuracy. By the continuous effort of so many researchers, deep learning algorithms are growing rapidly with an improved ob...
详细信息
ISBN:
(纸本)9781728151977;9781728151960
Object detection has seen many changes in algorithms to improve performance both on speed and accuracy. By the continuous effort of so many researchers, deep learning algorithms are growing rapidly with an improved object detection performance. various popular applications like pedestrian detection, medical imaging, robotics, self-driving cars, face detection, etc. reduces the efforts of humans in many areas. Due to the vast field and various state-of-the-art algorithms, it is a tedious task to cover all at once. This paper presents the fundamental overview of object detection methods by including two classes of object detectors. In two stage detector covered algorithms are RCNN, Fast RCNN, and Faster RCNN, whereas in one stage detector YOLO v1, v2, v3, and SSD are covered. Two stage detectors focus more on accuracy, whereas the primary concern of one stage detectors is speed. We will explain an improved YOLO version called YOLO v3-Tiny, and then its comparison with previous methods for detection and recognition of object is described graphically.
Waste management is a pressing global issue, and the need for efficient waste separation processes is becoming increasingly important. Incorporating Machine Learning techniques with waste separation has yielded promis...
Waste management is a pressing global issue, and the need for efficient waste separation processes is becoming increasingly important. Incorporating Machine Learning techniques with waste separation has yielded promising results and revolutionized the waste separation process. However, this research paper highlights the gaps that exist in the understanding and efficacy of machine learning for waste separation. This study provides an overview of state-of-the-art techniques for waste separation using machine learning, including the use of computer vision, sensor technology, and neural networks. This paper portrays case studies with empirical data from various waste management establishments to demonstrate the potential advantages of machine learning for recycling and waste classifying. Although progress has been made, this article highlights some significant gaps in our knowledge. These gaps include the need for standardized datasets, optimization of algorithms for real-world scenarios, the impact of technology on the environment, and the socio-economic aspects of integrating automation into waste management. The essential and most important tasks of intelligent mechanisms in the digital age is waste separation. Segregation of waste is necessary to apply proper disposal and waste management strategies. The existing systems use drones to identify waste by using deep learning, imageprocessing, GPS, and GSM approaches, locating and notifying the authorities about them. The improvement made is in the analysis and utilize multi-object detection and picture classification to accomplish trash segregation. Waste management may be effectively accomplished with higher accuracy through Machine Learning practices which therefore considerably lowers labor expenses.
暂无评论