International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
An Intelligent Intrusion Detection System Using Ensemble Learning for Ultra-Dense IoT Networks
Get PDFIntrusion detection refers to the process of observing and analyzing network or system incidents in a perpetual manner to identify unauthorized accesses, malicious acts, or violations of the rules. It plays a pivotal role in the protection of critical information, the prevention of security breaches, and the safety, confidentiality, and availability of company assets. Strong methods to identify and stop harmful activity are required because cybersecurity threats have grown more complex due to the quick expansion of digital infrastructure. Various researchers have conducted different research studies for intrusion detection, and different methodologies, along with traditional as well as machine learning models, have been applied with various datasets for the proposed task. This research aims to address these challenges by developing an efficient and intelligent intrusion detection system using a stacking ensemble learning approach. The proposed model integrates multiple base classifiers: Decision Tree, Naïve Bayes, K-Nearest Neighbor (KNN), and Linear Discriminant Analysis (LDA) to capture diverse decision boundaries, with a Random Forest acting as the meta-classifier to aggregate and optimize final predictions. The publicly available UNSW-NB15 dataset is employed in this study for intrusion detection. Python and its libraries are used for simulation purposes. After simulation, it has been achieved that the stacked model, which combines the predictions of multiple base learners through a meta-classifier, achieved a significantly higher accuracy of 99.93%. While in comparison, LDA achieved the highest accuracy of 94.25%, followed closely by SVM at 93.05%, DT at 91.00%, NB at 90.55%, and KNC at 89.81%. This demonstrates that ensemble learning, particularly stacking, can effectively leverage the strengths of individual models to greatly enhance intrusion detection performance for complex datasets
AI in the Field: A Review of Deep Learning Methods for Weed Identification in Wheat Crops
Get PDFWeed infestation is a major constraint in wheat production, causing yield losses and higher herbicide dependence. Traditional control methods often lack precision, highlighting the need for intelligent, sustainable solutions. Deep learning has recently emerged as a powerful tool for automated and accurate weed detection in precision agriculture. This review summarizes the latest advances in deep learning applied to wheat weed identification, emphasizing model architectures, datasets, and imaging techniques. Approaches such as YOLO variants, Faster R-CNN, U-Net, and transformer-based models have achieved high accuracy in distinguishing wheat from diverse weed species, even under complex field conditions. Integration of UAV imagery, multispectral sensors, and spectral indices further enhances detection at early growth stages. Recent innovations, including attention mechanisms, feature fusion, optimized loss functions, and lightweight designs, have improved precision, speed, and generalization. Key challenges remain in dataset quality, class imbalance, and cross-field applicability. This work outlines current trends, identifies gaps, and highlights future directions for scalable and sustainable deep learning-based weed detection in wheat agriculture
Effects of Polyethylene Glycol (PEG) Simulated Drought Stress on Physio-Agronomic Characteristics in Myhco Variety of Sorghum Bicolor L
Get PDFThe present study was aimed at determining the differential interactive effects of Ca/Mg quotient and PEG-simulated drought in Sorghum bicolor at the vegetative stage. Sorghum bicolor collected variety Myhco from Persabaq Nowshera were sown in earthen pots (lower inside diameter, 18cm upper inner diameter, 20 cm height and 2 cm thickness) filled with 2 kg of air-dried soil and silt (2:1) pH, moisture content and field capacity in triplicates in the green house of the Department of Botany, University of Peshawar in 2019. The designed experiment contains seven treatments each having three replicates, among these treatments first three are control, the second three are treated Ca/Mg quotient 4+PEG0.6 Ca/Mg quotient 4+PEG0.2, Ca/Mg quotient 2+PEG0.6, Ca/Mg quotient 2+PEG0.2, Ca/Mg quotient 0.18+PEG0.6 Ca/Mg quotient 0.18+PEG0.2, while the last three treatments are treated Ca/Mg quotient 0.18+PEG0.2. Conclusions We conclude that there is a reduction in the agronomy, i.e., leaf area, leaf fresh and dry weight, and a similar reduction also occurred with all other vegetative parts. There is a clear difference between control and PEG drought, and a greater reduction is observed in 0.6 MPa drought. The biochemical characters were also affected in the same manner; a clear reduction was observed in chlorophyll, sugar, and protein, and occurred while the Ca/Mg quotient had no significant effect on Sorghum bicolor L.in Varity myco
Artificial Intelligence-Augmented Intrusion Detection Systems for Advanced Threat Taxonomy in Cloud Computing Environments
Get PDFOver the past few decades, cyber-attacks have emerged as a grave form of criminal activity and a subject of intense scholarly and policy debate. The rapid proliferation of cloud computing services— particularly Software as a Service (SaaS)—has further motivated research to classify security threats and their corresponding countermeasures. Scholars have increasingly focused on the risks, vulnerabilities, and malicious intrusions inherent in such environments, with particular emphasis on MITM (MITM) attacks and their mitigation and detection mechanisms. Host-based virtual software has demonstrated considerable efficacy in detecting malware within localized environments. Building on this foundation, the present study classifies Man-in-the-Middle (MITM) attacks in SaaS platforms through the deployment of Cloud-based Intrusion Detection Systems (CIDS). Our investigation concentrates specifically on attacks that target cloud hosts deployed within SaaS infrastructures. The proposed methodology incorporates the roles of the source cloud, destination cloud, and directional flow of the attack vector. In this context, the cloud ecosystem is understood as a dynamic environment where any participating entity, equipped with sufficient technical expertise, may both launch and be subjected to sophisticated intrusions. Accordingly, adaptive CIDS monitoring architectures are essential to safeguard communication between cloud actors. Moreover, CIDS frameworks furnish modular components capable of aggregating alerts, conducting analysis, and notifying administrators of potential breaches. To further illustrate the threat landscape, we present a statistical analysis of vulnerabilities most frequently exploited in MITM scenarios. This classification not only highlights the evolving tactics of adversaries but also equips readers with a structured understanding of MITM attacks, thereby fostering greater familiarity with contemporary cloud security challenges
Data Mining for Smarter Administration of TVET Institutes
Get PDFRetaining the trainees is a major problem for the TVET institutes today. The trend of TVET education in Khyber Pakhtunkhwa province has improved in recent decades. Despite its high cultural barriers, resistance in women\u27s education, and dropout rates, on the basis of annual admission, Khyber Pakhtunkhwa holds the 2nd position among all other provinces of Pakistan. In this research, we have tried to decrease the dropout ratio by enhancing the Daily attendance of the trainees and improving their results. Monthly Fee slip, Date Sheet, and Results will be shared with the parents/guardians through SMS/WhatsApp. New TVET institutes will be able to check the trainee\u27s educational record from the previous TVET institute. The Data Mining for Smarter Administration of TVET Institutes will be a Mobile and Web-Based Application and will keep a close relationship between Parents, Teachers, and Administration of the TVET institute
Design of a Miniaturized Flexible Patch Antenna with Shorting-Pin Integration for Enhanced Gain in RFID, ISM, and Wearable Biomedical Applications
Get PDFAntennas are integral parts of wireless communication because they can ensure that signals are transmitted and received effectively, encompassing a variety of frequencies, such as those used in IoT and overall RF systems. This paper introduces a miniaturized single-band antenna of 9 × 31.6 × 0.254 mm3 with a specific design for wearable applications, which was made on flexible Rogers RT5880 substrate. Using the CST Microwave Studio 2024 microchip as the design and analysis tool, the proposed antenna consists of a rectangular slot-based radiating structure with a shorting pin and probe feed, which can achieve stable design performances both in the free-space and on-body situation services. Compliant with the wearable safety requirements and operating at a low level specific absorption rate (SAR) less than 1.6 W/kg at the resonant frequency. The antenna has good radiation performance, with a maximum efficiency of 82% and a maximum gain of 4.1 dBi, having a bidirectional radiation pattern in the elevation plane and an omnidirectional radiation pattern in the azimuth plane. The introduction of shorting-pin as a strategic method of reducing the length of resonance allows the approach of substantial reduction in the resonant length without compromising radiation characteristics. Simulation results also provide further confidence scales of stabilizing impedance performance and omnidirectional radiation pattern properties in the target 5.725-5.875 GHz ISM band, which demonstrates the expectation of shorting-pin strategies for the development of small-size, high biocompatibility, and flexible antennas for next-generation wearable, body-area network, and radio frequency identification (RFID) communication systems
Exploring the Influence of Teacher\u27s Stress and Emotions on Student Behavior in Secondary Schools
Get PDFThe focus of this study is on the intricate relationship between the stress of the teacher, the emotion of the teacher, and the behavior of the workers in secondary school classrooms, stressing the bidirectional relation (between) the emotional well-being of (the teacher) and the engagement of (the worker). Qualitative data from interviews with teachers and students are drawn on to investigate how emotional exhaustion, workload pressures, and the absence of institutional support prevent teacher regulation of emotion, and inhibit effective classroom management. What they found was that students can pick up on teacher stress, often based on what students perceive to be nonverbal cues, tone, and expressions, and those interpretations lead to the classroom behaviors of the students. Positive emotions by teachers develop trust, focus, and engagement, whereas frustration, inconsistency, and unpredictability engender student disengagement, increased disruptive behaviors, and strained relationships between teachers and students. These same emotional dynamics are catalyzed further by cultural factors; for instance, hierarchical power structures influence the context within which students respond to teacher stress. The systemic institutional support mechanisms are also lacking resulting in a cascade of emotional exhaustion followed by diminished classroom effectiveness. This study highlights the importance of emotional intelligence training, proactive classroom management strategies, and robust institutional support systems in a bid to reduce teacher stress and create a stable, positive learning environment. These results add to existing literature on the emotional well-being of teachers and provide practical suggestions to educational policymakers, administrators, and teacher training programs on addressing the emotional and psychological demands of teachers in today’s classrooms
Steering Control of Ackermann Architecture Weed Managing Mobile Robot
Get PDFA robot that finds and eliminates weeds from crops is called a weed control robot. Weeds deplete primary crops moisture supplies and hinder their development. They may be harmful to both human and animal health and result in losses in crop yield. Herbicides and other chemicals have been used for many years to eradicate weeds from crops; nevertheless, these chemicals harm plants and contaminate the environment. In this work, a novel semantic weeds detection approach based on PC/BC-DIM network has developed which shows outbreaking performance and classification results compared to the state-of-art approaches. We developed an autonomous weed control robot which consists of Ackermann Architecture and delta robot. Delta robot have a camera on its base that is used to detect the real time weeds in the environment. First of all, image is acquired by camera and with the help of image processing techniques we are able of detecting the weed from other crops and eliminate them by the electrical discharging method in which electrodes are connected at its end effector that will burn the weed detected. We also developed a system for path planning and obstacle avoidance for navigation of mobile robot in which we used the technique of stereo vision that will capture the stereo images of environment and find their disparity. With the help of depth information, robot will be able to detect the object in its way and avoids the obstacle and find the shortest path to navigate in field using A* algorithm. The results obtained from this work are simulation based which are detection of weed in field images using image processing and path planning of robot using stereo images of field. The system has a fairly good overall accuracy of 81.25%. The efficiency of the system is moderate, but the relatively high False Positive Rate and RMS Error suggest that the system need improvement to reduce significant errors and false positives. Our future work involves the removal of weed and implementation of simulated results to hardware
Cow Face Detection for Precision Livestock Management using YOLOv8
Get PDFPrecision livestock management is transforming traditional agricultural practices by boosting productivity, increasing yield, and automating tasks, all while reducing labor requirements and minimizing errors. Conventional methods for animal recognition are often unreliable, which has led to a growing preference for using cameras to identify animals, monitor their health, manage data, and maintain cattle records. However, small-scale farms with limited livestock, such as cows and goats, frequently face overfitting problems in traditional machine learning models due to insufficient training data. Identifying individual cows based on facial features becomes more effective after detecting the cow’s face. This study addresses these challenges by fine-tuning YOLOv8, a pretrained model, using a mix of self-captured images and publicly available datasets to detect cow faces in complex environments. Integrating publicly available data and leveraging a pretrained COCO model has significantly improved the model’s ability to generalize and accurately detect cow faces. YOLOv8, equipped with the COCO pretrained model, successfully detects nearly all types of cow faces, which can then be used for individual cow classification. This approach enhances cow recognition accuracy, contributing to more efficient farm management applications
Efficient Region-Based Video Text Extraction Using Advanced Detection and Recognition Models
Get PDFThis paper presents an automated process for extracting text from video frames by specifically targeting text-rich regions, identified through advanced scene text detection methods. Unlike traditional techniques that apply OCR to entire frames—resulting in excessive computations and higher error rates—our approach focuses only on textual areas, improving both speed and accuracy. The system integrates effective preprocessing routines, cutting-edge text detectors (CRAFT, DBNet), and advanced recognition engines (CRNN, transformer-based) within a unified framework. Extensive testing on datasets such as ICDAR 2015, ICDAR 2017 MLT, and COCO-Text demonstrates consistent gains in F-scores and word recognition rates, significantly outperforming baseline methods. Additionally, detailed error analysis, ablation studies, and runtime evaluations offer deeper insights into the strengths and limitations of the proposed method. This pipeline is particularly useful for tasks like video indexing, semantic retrieval, and real-time multimedia analysis