Browse Items (11810 total)
Sort by:
-
Detection of Disease in Mango Trees Using Color Features of Leaves
The goal has been to detect disease in mango trees. This paper compares different approaches to extract color features and check the accuracy and applicability for mango trees. The paper proposes variations which helped in increasing the accuracy of features extracted for mango trees: firstly, a customized method of splitting leaf into layers while doing K-means clustering, and secondly, segmenting the region of interest to blocks to help in applying statistical functions more accurately over a region. 2020, Springer Nature Singapore Pte Ltd. -
Detection of DoS Attacks Using Machine Learning Based Intrusion Detection System
Conventional intrusion detection systems are not always sufficient due to the increasing sophistication and frequency of Denial-of-Service (DoS) attacks. This work presents a novel solution to this problem by leveraging machine learning techniques to increase the precision and efficacy of real-time intrusion detection. The system keeps a careful eye on network traffic patterns, looking for any irregularities that would point to a denial-of-service attack. An Intrusion Detection System (IDS) that utilizes machine learning technologies - specifically, neural networks and support vector machines - allows for real-time adaptation to new attack patterns. A combination of rigorous simulations and real-world testing provides empirical support for the IDS's quick detection and mitigation of DoS threats. This initiative makes a major contribution to the development of cybersecurity defenses. 2024 IEEE. -
Detection of faces from video files with different file formats
Face detection is the primary approach of all fundamental problems of human computer interaction system (HCIS). This paper evaluates the performance of detection system on single face from stored videos that are stored in different file formats. Stored videos contain raw homemade datasets as well as ready-made datasets. This proposed work concludes detection percentage of face detection system in different video formats. The implementation is done in two phases. The raw homemade dataset is tested on.3gp,.avi,.mov,.mp4 and a ready-made dataset is tested on.wmv,.m4v,.asf,.mpg file formats. The coding part for face detection has been done in MATLAB R2013a. The detection of faces from video file was 72.79 % for homemade dataset and 82.78% for ready-made dataset. 2016 IEEE. -
Detection of fake opinions on online products using decision tree and information gain
Online reviews are one of the major factors for the customers to purchase any product or to get service from many sources of information that can be used to determine the public opinion on the products. Fake reviews will be published intentionally to drive the web traffic towards the particular products. These fake reviewers mislead the customers to distract the purchasers mind. Reviewers behaviors are extracted based the semantical analysis of his review content for the purpose of identifying the review as fake or not. In this work the reviews are extracted from the web for a particular product, along with the reviews of several other information related to the reviewers also been extracted to identify the fake reviewers using decision tree classifier and Information Gain.Significance of the features on the decision is validated using information gain. Experiments are conducted on exhaustive set of reviews extracted from the web and demonstrated the efficacy of the proposed approach. 2019 IEEE -
Detection of Forest Fire Using Modified LSTM Based Feature Extraction with Waterwheel Plant Optimisation Algorithm Based VAE-GAN Model
A crucial natural resource that directly affects the ecology is forests. Forest fires have become a noteworthy problem recently as a result of both natural and man-made climatic changes. A smart city application that uses a forest fire discovery technology based on artificial intelligence is provided in order to prevent significant catastrophes. A major danger to the environment, animals, and human lives is posed by forest fires. The early detection and suppression of these fires is crucial. This work offers a thorough method for detecting forest fires using advanced deep learning (DL) algorithms. Preprocessing the forest fire dataset is the initial step in order to improve its relevance and quality. Then, to enable the model to capture the dynamic character of forest fire data, long short-term memory (LSTM) networks are used to extract useful feature from the dataset. In this work, weight optimisation in LSTM is performed using a Modified Firefly Algorithm (MFFA), which enhances the model's performance and convergence. The Variational Autoencoder Generative Adversarial Networks (VAEGAN) model is used to classify the retrieved features. Furthermore, every DL model's success depends heavily on hyperparameter optimisation. The hyperparameters of an VAEGAN model are tuned in this research using the Waterwheel Plant Optimisation Algorithm (WWPA), an optimisation technique inspired by nature. WPPA uses the idea of plant growth to properly tune the VAEGAN's parameters, assuring the network's peak fire detection performance. The outstanding accuracy (ACC) of 97.8%, precision (PR) of 97.7%, recall (RC) of 96.26%, F1-score (F1) of 97.3%, and specificity (SPEC) of 97.5% of the suggested model beats all other existing models, which is probably owing to its improved architecture and training techniques. Copyright: 2024 The authors. This piece is published by IIETA and is approved under the CC BY 4.0 license. -
Detection of Fraudulent Alteration of Bank Cheques Using Image Processing Techniques
In todays world illegal alteration and illegal modifications of authenticated financial documents is increasing rapidly as a fastest growing crimes around the world. The result of this kind of crimes may result in a huge financial loss. In this paper image processing and document image analysis techniques are used to examine such cases in order to identify the fraudulent bank cheques. However, it is very difficult to detect an alteration made on documents once the printing ink of alike color is employed. In this paper, alterations and modifications caused with handwritten ball point pen strokes are considered and proposed a technique for recognition of such types of corrections by employing standard techniques under Digital image processing and pattern recognition. The results are quite promising during the experiments conducted. 2021, Springer Nature Singapore Pte Ltd. -
Detection of high-frequency pulsation in WR135: Investigation of stellar wind dynamics
We report the detection of high-frequency pulsations in WR 135 from short-cadence (10 minute) optical photometric and spectroscopic time series surveys. The harmonics up to the sixth order are detected from the integrated photometric flux variations, while the comparatively weaker eighth harmonic is detected from the strengths of the emission lines. We investigate the driving source of the stratified winds of WR 135 using the radiative transfer modeling code, CMFGEN, and find the physical conditions that can explain the propagation of such pulsations. From our study, we find that the optically thick subsonic layers of the atmosphere are close to the Eddington limit and are launched by the Fe opacity. The outer optically thin supersonic winds (Tross = 0.1 0.01) are launched by the He II and C IV opacities. The stratified winds above the sonic point undergo velocity perturbation that can lead to clumps. In the optically thin supersonic winds, dense clumps of smaller size (fVFF = 0.27 0.3, where fVFF is the volume filling factor) pulsate with higher-order harmonics. The larger clumps (fVFF = 0.2) oscillate with lower-order harmonics of the pulsation and affect the overall wind variability. 2024. The Author(s). -
Detection of Lung Cancer with a Deep Learning Hybrid Classifier
This article presents a deep learning framework combining a convolutional neural network (CNN) and a support vector machine (SVM) for lung cancer diagnosis. The model uses data divided into six groups: 250 images in the training set and 150 images in the test set. The work includes preliminary data and development using the Keras image data generator, VGG-16 architecture, high-level rules, and SVM classifier training with labels and vectors. The model achieves 90% accuracy with 85% selection impact and 75% cross-validation flexibility using VGG-16 and SVM hybrid classifier. This study finally revealed the classification of the model by multi-class ROC curve analysis and confusion matrix. 2024 IEEE. -
Detection of Malicious Nodes in Flying Ad-hoc Network with Supervised Machine Learning
An Ad-hoc network (FANET) is a new upcoming technology which has been used in several sectors. Ad-hoc networks are mostly wireless local area networks (LANs). The devices communicate with each other directly instead of relying on a base station or access points as in wireless LANs for data transfer. In an Ad-hoc network the communication between one node to another in a FANET is not secured and there isn't any authorized protocol for secured communication. Therefore, we suggest an algorithm to detect the malicious node in a network. This algorithm uses Linear regression to calculate the reputation or trust value of a node in the network. Then the above found trust value is used to classify the node as normal node or malicious node based on the Logistic Regression Classification. Thus, allowing a secure communication of data and avoiding attacks. 2022 IEEE. -
Detection of picric acid in industrial effluents using multifunctional green fluorescent B/N-carbon quantum dots
Carbon quantum dots have recently gained widespread attention due to their excellent physicochemical features. The rapid escalation in the dumping of hazardous chemicals into water, spurred demand for developing efficient and selective sensors for toxic chemicals. Herein, we have developed a novel fluorescence sensor for picric acid which is a major pollutant in industrial effluents. The new strategy exploits the development of a fluorescence sensor based on N-doped carbon quantum dots (N-CQDs) followed by boron functionalization. The N-CQDs were synthesized in a rapid single-step microwave technique by employing L-serine and citric acid. Subsequent boron functionalization of N-CQDs was carried out using boric acid for the synthesis of Boron-nitrogen carbon quantum dots (B/N-CQDs). The B/N-CQDs were found to exhibit high quantum yield (24%), good water solubility, outstanding photostability features, and bright green fluorescence under UV light. The morphology of B/N-CQDs is spherical, with scattered particle sizes ranging from 2 to 8 nanometers. Furthermore, B/N-CQDs were found to be an effective fluorescence probe for the selective and sensitive detection of picric acid, with a good linear range of 37 nM-30 M and a detection limit of 1.8 nM. The Photoluminescence (PL) intensity of B/N-CQDs was selectively quenched by picric acid. The quenching mechanism was conclusively established using fluorescence lifetime decay studies. Moreover, the synthesized B/N-CQDs was successfully employed for the analysis of picric acid from industrial effluents and cell imaging with Hela cells to showcase the utility of the developed fluorescent probe. 2022 Elsevier Ltd -
Detection of picric acid in industrial effluents using multifunctional green fluorescent B/N-carbon quantum dots /
Journal of Environmental Chemical Engineering, Vol.10, Issue 2, ISSN No: 2213-3437.
Carbon quantum dots have recently gained widespread attention due to their excellent physicochemical features. The rapid escalation in the dumping of hazardous chemicals into water, spurred demand for developing efficient and selective sensors for toxic chemicals. Herein, we have developed a novel fluorescence sensor for picric acid which is a major pollutant in industrial effluents. The new strategy exploits the development of a fluorescence sensor based on N-doped carbon quantum dots (N-CQDs) followed by boron functionalization. The N-CQDs were synthesized in a rapid single-step microwave technique by employing L-serine and citric acid. -
Detection of strangers based on dogs sound
Nowadays, people having a pet at home are increasing. Usually, dog is the favorite pet animal for most of the people in the world. Dogs are more capable of identifying strangers in the surroundings than humans. The proposed work identifies the strangers based on the barking sound of the dog. In this anticipated work, multiple features are extracted from the dogs barking sound using Fast Fourier Transform and Statistical based methods. The classification is done using Nae Bayes classifier. The dataset contains 770 barking audio files of 8 dogs. Whenever known and unknown person comes home, the sounds of the dogs are recorded. The classification result for identifying the stranger is 79.1094%. BEIESP. -
Detection of toxic comments over the internet using deep learning methods
People now share their ideas on a wide range of topics on social media, which has become an integral part of contemporary culture. The majority of people are increasingly turning to social media as a necessity, and there are numerous incidents of social media addiction that have been reported. Socialmedia channels. Socialmedia platforms have established their worth over time by bringing individuals from different backgrounds together, but they have also shown harmful side effects that could have serious consequences. One such unfavourable result is how extremely poisonous many discussions on social media are. Online abuse, hate speech, and occasionally outrage culture are now all considered to be toxic. In this study, we leverage the Transformers Bidirectional Encoder Representations to build an efficient model to detect and classify toxicity in user-generated content on social media. The Kaggle dataset with labelled toxic comments, was used to refine the BERT pre-trained model. Other Deep learning models, including Bidirectional LSTM, Bidirectional-LSTM with attention, and a few other models, were also tested to see which performed best in this classification task. We further evaluate the proposed models utilising dataset obtained from Twitter in order to find harmful content (tweets) using relevant hashtags. The findings showed how well the suggested methodology classified and analysed toxic comments. 2024 selection and editorial matter, Arvind Dagur, Karan Singh, Pawan Singh Mehra & Dhirendra Kumar Shukla; individual chapters, the contributors. -
Detection of toxic comments over the internet using deep learning methods
People now share their ideas on a wide range of topics on social media, which has become an integral part of contemporary culture. The majority of people are increasingly turning to social media as a necessity, and there are numerous incidents of social media addiction that have been reported. Socialmedia channels. Socialmedia platforms have established their worth over time by bringing individuals from different backgrounds together, but they have also shown harmful side effects that could have serious consequences. One such unfavourable result is how extremely poisonous many discussions on social media are. Online abuse, hate speech, and occasionally outrage culture are now all considered to be toxic. In this study, we leverage the Transformers Bidirectional Encoder Representations to build an efficient model to detect and classify toxicity in user-generated content on social media. The Kaggle dataset with labelled toxic comments, was used to refine the BERT pre-trained model. Other Deep learning models, including Bidirectional LSTM, Bidirectional-LSTM with attention, and a few other models, were also tested to see which performed best in this classification task. We further evaluate the proposed models utilising dataset obtained from Twitter in order to find harmful content (tweets) using relevant hashtags. The findings showed how well the suggested methodology classified and analysed toxic comments. 2024 The Author(s). -
Detection of tuberculosis using convolutional neural network with transfer learning
Tuberculosis is sighted as the one of the life causing disease in the recent time. The current research work focus on detection of Tuberculosis using Convolutional Neural Network with Transfer Learning for chest X-ray images. The proposed research work uses two different datasets for detecting Tuberculosis from Chest X-ray images, which is taken from National Institutes of Heaths. During the experimental work, the total sample size used for detecting Tuberculosis is 800 instances. Initially, the image processing techniques were applied to increase the quality of Chest X-ray images. The proposed model uses Convolution Neural Network with transfer learning for the detection of Tuberculosis with 98.7% as accuracy. The proposed model is checked with convolutional neural network without transfer learning. From the experimental evaluation, it is found that the proposed model works better than the Convolution Neural Network without using the transfer learning. 2017, Institute of Advanced Scientific Research, Inc. All rights reserved. -
Detection of Various Security Threats in IoT and Cloud Computing using Machine Learning
Due to the growth of internet technology, there is a sharp rise in the growth of IoT enabled devices. IoT (Internet of Things) refers to the connection of various embedded devices with limited processing and memory. With the heavy adoption of IoT applications, cloud computing is gaining traction with the ever-increasing demand to process and compute a massive amount of data coming from various devices. Hence, cloud computing and IoT are often related to each other. However, there are two challenges in deploying the IoT and cloud computing frameworks: security and Privacy. This article discusses various types of security threats affecting IoT and cloud computing, and threats are classified using machine learning (ML). ML has gained much momentum in recent years and is applied in various domains. One of the main subdomains of machine learning is used in IoT and cloud security. A machine learning model can be trained with data based on which the model can predict the impending security threats. Popular security techniques to protect IoT devices from hackers are IoT authentication, access control, malware detection, and secure overloading. Supervised learning algorithms can be used to detect malware in the runtime behavior of applications. The malware is detected from network traffic and is labeled based on its suspicious behavior. Post identification of malware, the application data is stored in a database trained via an ML classifier algorithm (KNN or Random Forest). With increased training, the model can identify malware applications with higher accuracy. 2022 IEEE. -
Detection of X-ray polarization in the high synchrotron peaked blazar 1ES 1959+650
We report the measurement of X-ray polarization in the high synchrotron peaked blazar 1ES 1959+650. Of the four epochs of observations from the Imaging X-ray Polarimetry Explorer, we detected polarization in the 28 keV band on two epochs. From the model-independent analysis of the observations on 28 October 2022, in the 28 keV band, we found the degree of polarization of ?X=9.01.6% and an electric vector position angle of ?X=535 deg. Similarly, from the observations on 14 August 2023, we found ?X and ?X values as 12.50.7% and 202 deg, respectively. These values are also in agreement with the values obtained from spectro-polarimetric analysis of the I, Q, and U spectra. The measured X-ray polarization is larger than the reported optical values, ranging between 2.5% and 9% when observed from 2008 to 2018. Broadband spectral energy distribution constructed for the two epochs is well described by the one-zone leptonic emission model with the bulk Lorentz factor (?) of the jet larger on 14 August 2023 compared to 28 October 2022. Our results favor the shock acceleration of the particles in the jet, with the difference in ?X between the two epochs being influenced by a change in the ? of the jet. Indian Academy of Sciences 2024. -
Determinant of Capital Structure in Indian Manufacturing Sector
Asia-Pacific Journal of Management Research and Innovation Vol. 8, No. 3. pp 265-269, ISSN No. 2319-510X -
Determinants and Impacts of Mergers and Acquisitions in the Drugs and Pharmaceutical Industry in India
Mergers and Acquisitions (MandAs) are inorganic growth strategies adopted by firms for achieving the objective of long-term growth maximization. Compared to other inorganic growth strategies like joint ventures and strategic alliances, MandAs offer deeper restructuring opportunities and better control over business over a long-term newlinebasis. During the third wave of globalization which started in early 1990s, MandAs became a popular strategy for firms to expand their businesses beyond the national boundaries. newlineIndian economy has been witnessing buzzling activity in the MandA landscape. A sectoral analysis of MandA trends identifies pharmaceutical sector as one of the top 5 newlinesectors with the highest MandA deal values during the period 2013-2016. Though Pharma sector has witnessed a decline in deal values during few years in the recent past, the resilience of the sector is visible through its ability to bounce back with record newlinebreaking deal values. Due to the continuous regulatory changes occurring in the domestic and foreign markets, pharma companies have to constantly change their strategies to survive and grow in the industry. MandAs enable pharma companies to adapt to these changes quickly. This study explores how the firms in the pharmaceutical sector use MandA as a strategy to navigate through the dynamic competitive landscape. The objectives of this research are threefold developing an understanding of the motives behind MandA decisions of the pharma firms, identification of the firm level determinants of acquisition probability and assessment of impact of MandAs. This study newlineuses qualitative content analysis for identification of MandA motives. The firm level newlinedeterminants of acquisition probability have been explored using Random Effect Logistic (REL) regression using panel data. Case study approach has been employed to assess the MandA impact by comparing the MandA motives with the post MandA outcomes. -
Determinants of adoption of digital payment services among small fixed retail stores in Bangalore, India
India is well on its way to becoming a trillion-dollar digital economy and the government is actively working towards it. Digital payment is taking up and gaining momentum in India. Digital payments have penetrated in all parts of life in India. But it is reported that digital payments are less penetrated among small vendors across the country. This study intends to identify and analyse the factors that determine the adoption of digital payment technologies among small fixed retail stores in tier 1 cities such as Bangalore. The study is based on primary data which is collected through well-structured questionnaires from small fixed retail merchants. The collected data are analysed to determine the factors affecting the adoption of digital payment services among small fixed retail merchants using appropriate statistical tools. The study has found that habit, pervasiveness, and operating costs are the factors that significantly affect the adoption of digital payment services among small fixed retail merchants. Copyright 2022 Inderscience Enterprises Ltd.