International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
A Dynamic Architecture to Control Multi-Rotors Using Hand Gestures
Get PDFTraditional methods for controlling multi-rotors typically involve joysticks, radio controllers, and mobile applications. However, these methods pose significant challenges, particularly for novice users like farmers, due to the extensive training and understanding required to effectively operate a copter. This paper introduces a highly adaptable architecture designed to offer an end-to-end solution for controlling a copter using hand gestures. The proposed system leverages a depth sensor and Convolutional Neural Network (CNN) to recognize hand gestures, utilizing a custom dataset collected from both indoor and outdoor environments. Through a series of simulations with novice users, the system has demonstrated successful operation in real-world scenarios. Currently, the architecture can accurately recognize six distinct gestures with an average accuracy of 90.5% across three different test environments with varying lighting conditions. Key features of this proposed solution include its adaptability, reliable performance, especially in low-light conditions, and its user-friendly design, making it particularly well-suited for farmers and other inexperienced users
Enhancing Management Strategies: Machine Learning and Creative Performance Insights in Employee Attrition Analysis and Prediction
Get PDFEmployee attrition and excessive turnover are major difficulties in today\u27s competitive employment market, affecting many industries. To overcome these difficulties, firms are increasingly relying on artificial intelligence (AI) to forecast staff loss and devise effective retention strategies. This study investigates famous machine learning (ML) models to forecast employee turnover and deliver data-driven solutions. The first section of the study compares various ML models on an imbalanced dataset. The second section introduces the Synthetic Minority Oversampling Technique (SMOTE) for data oversampling and applies ML models to the enlarged dataset. ML can predict employee turnover by examining historical data, employee behavior, and external factors. This early detection enables organizations to respond proactively with targeted retention strategies. The study concludes that the Random Forest model is the best model when combined with SMOTE, achieving performance scores of 0.96 out of 1
Analyzing the Impacts of Soapstone Dust on Respiratory System of Mine Workers Through Structural Equation Modelling Technique: A Case Study of Sherwan Soapstone Mines, Abbottabad, Pakistan
Get PDFDust produced in mining has a substantial impact on worker’s health resulting in severe respiratory diseases. Researchers mainly focused on the dust problems faced in surface mining whereas the dust produced in underground soapstone mines has received comparatively less attention. This study evaluates self-reported respiratory symptoms and medical examinations of underground mine workers in soapstone mines. It establishes a relationship between the respiratory illness factors and its symptoms, providing new insight into the analysis. Demographic and other respiratory symptoms-related data is collected through questionnaires from underground soapstone mine workers, located in the Abbottabad area, with medical data from 60 of these workers obtained through medical examinations. The collected data is subsequently analyzed using Structural Equation Modelling and regression analysis to investigate the relationship between the evaluated factors in the dust analysis. The dust assessment shows that it is primarily composed of silica, with small particle sizes that are smaller than the threshold limit value and pose a risk of silicosis. The questionnaire data indicates that about 75% of workers exhibit symptoms of respiratory diseases, the majority of them are laborers and old age workers whereas the medical examinations revealed that 80% of workers are affected by lung infections. The Structural Equation Modelling demonstrates that dust inhalation has a stronger effect on symptom occurrence (β = 0.485, p < 0.001) compared to dust severity (β = 0.207, p < 0.05). These results are concerning and underscore the need for interventions, and the adoption of adequate respiratory protection measures for safeguarding the health of workers
Image Compression Exploration using Discrete Wavelets Transform Families and Level
Get PDFThis analysis paper is based on Discrete Wavelets Transform (DWT) for image compression using wavelets families and levels. The DW transforms the image or data into frequency components that match its resolution scale while the compression removes duplication and unwanted information on the receiver side. Wavelets in compression observe the whole image very finely and thus produce no blocking artifacts. Thus, wavelets are high-quality image compression used in many real-world applications i.e. image, multimedia, biometric and biological analysis, computer graphics and image processing, etc. In this investigation, first of all, various compression methods have been compared. It is validated based on compression ratio that DWT is the optimal choice. Secondly, for experimental and analysis purposes, random real-time digital images both RGB and greyscale have been used as a dataset. The assessment images have been converted to grayscale if RGB, decomposed using wavelet levels, and compressed using wavelet families. Threshold coefficients have been evaluated by the Birge-Massart strategy using two scenarios i.e. simulator control thresholding and increasing threshold. Birge-Massart thresholding is best for the compression of still images in wavelet transform. e evaluation and comparison of various wavelet families and decomposition levels were conducted based on criteria such as image compression effectiveness, retained energy, and zero coefficients. The size of original, compressed, and decompressed images has also been computed and displayed for analysis purposes. The analysis of wavelet families and decomposition levels indicated that increasing levels up to a certain range for decomposition purposes in various wavelet compression families enhances image smoothness consistently. With image smoothness, roughness, and noise spikes in images have been reduced. However, it is observed that after specific levels, image quality degradation has been observed. The significance and novelty of the work provide analysis for appropriate and effective quality image compression using DWT families and levels in different applications. The purpose is to reduce need-based storage requirements and lightweight transmission. Additionally, the optimum compression algorithm in DWT families and levels is also found based on the results. As selection of wavelet filters and decomposition level play an important role in achieving an effective compression performance because no filter performs the best for all images
Optimized Coverage and Capacity Planning of Wi-Fi Network based on Radio Frequency Modeling & Propagation Simulation
Get PDFInvestigation for optimized coverage and capacity planning of Wi-Fi network is carried out in the testbed for the purpose of optimization in terms of Received Signal Strength Indicator (RSSI), Signal to Noise & Interference Ratio (SNIR), Interference + Noise (I+N), downlink/uplink data rate and user capacity. The plan is carried out by conducting a site prediction survey through Altair’s Win Prop Software which is a Radio Frequency (RF) modeling and signal propagation simulation software, using the configuration of actual Wireless Local Area Network-Access Points (WLAN-APs). First, the map of the testbed with all respective material properties is drawn in Win Prop’s Wall Manager (Wall Man) Tool as a 3-Dimentional (3D) model. Then that 3D model is implemented in Win Prop’s Propagation Manager (ProMan) Tool where APs are deployed and wave propagation analysis as well as capacity planning is done. Results are analyzed for optimal signal strength, data rate, and user handling capacity. The results are validated by a smartphone-embedded software known as Cellular-Z. The average optimization increase in coverage, downlink & uplink data rates is 3.95 dB, 2.53 Mbps & 3.42 Mbps respectively
Nature Scene Classification Using Transfer Learning with Inception V3 on the Intel Scene Dataset
Get PDFNature scene classification is vital for various applications, including environmental monitoring and autonomous systems need to develop efficient models that can sort out different scenes. This work proposes a new approach using state-of-the-art CNNs like InceptionV3, Xception and VGG19, to enhance the classification accuracy and generalization of nature scenes. We worked with six classes with 20,926 training images and 5,228 validation images and augmented the data to improve the model. Models were fine-tuned from the pre-trained models of ImageNet and early stopping and model checkpoints were used to avoid overfitting. The results indicated that the proposed InceptionV3 model achieved a training accuracy of 94.49% and validation accuracy of 92.81% which is higher than previous work and Xception model had a high accuracy of 95.52% but the model might be overfitting. During the comparison of the results, it was revealed that InceptionV3 provided the highest accuracy with the least standard deviation, which proved the effectiveness of the selected architecture for scene classification. These results indicate that the selection of the model and the technique for the classification of nature scenes is important. It is a good advancement in the field of nature scene classification and provides a reliable solution to enhance accuracy in real-world scenarios
Optimizing Human Activity Recognition with Ensemble Deep Learning on Wearable Sensor Data
Get PDFIn recent years, the research community has shown a growing interest in the continuous temporal data gathered from motion sensors integrated into wearable devices. This type of data is highly valuable for analyzing human activities in a variety of domains, including surveillance, healthcare, and sports. Various deep-learning models have been developed to extract meaningful feature representations from temporal sensory data. Nonetheless, many of these models are constrained by their focus on a single aspect of the data, frequently overlooking the complex relationships between patterns. This paper presents an ensemble model aimed at capturing these intricate patterns by combining CNN and LSTM models within an ensemble framework. The ensemble approach involves combining multiple independent models to harness their strengths, resulting in a more robust and effective solution. The proposed model utilizes the complementary capabilities of CNNs and LSTMs to identify both spatial and temporal features in raw sensory data. A comprehensive evaluation of the model is conducted using two well-known benchmark datasets: UCI-HAR and WISDM. The proposed model attained notable recognition accuracies of 97.92% on the UCI-HAR dataset and 98.52% on the WISDM dataset. When compared to existing state-of-the-art methods, the ensemble model exhibited superior performance and effectiveness
The Implementation of Rent Hub System: An Intelligent Online Rental Marketplace with ML-Powered Personalized Product Discovery and Recommendations
No full textThe rapid expansion of peer-to-peer rental services has significantly influenced the share economy by connecting consumers with short-term access to diverse rental products. However, existing platforms primarily focus on specific categories, limiting consumer choices and creating a gap in the market. This study introduces RentAll, a comprehensive multi-category rental platform offering access to houses, automobiles, furniture, gadgets, and jewelry, while prioritizing data privacy through anonymized transactions. To enhance user experience, we developed a recommendation system utilizing content-based filtering, cosine similarity, and collaborative filtering through FP-Growth Frequent Itemset Mining to suggest products based on customer behavior. Additionally, a chatbot powered by a Sequence-to-Sequence model using RNN and LSTM units was integrated for real-time customer support. The results demonstrate RentAll\u27s effectiveness in providing a unified rental solution with personalized recommendations. The platform streamlines the rental process, reduces financial strain, and expands product offerings to serve diverse demographics. High user satisfaction is reported due to its user-friendly interface and engaging features, including secure payment processing via Easypaisa. Moreover, the implementation of robust security measures protects user information and builds trust. In conclusion, RentAll effectively addresses key issues in online rentals by offering a user-friendly platform with diverse rental categories, enhancing consumer convenience and satisfaction while maintaining stringent data protection standards
Development of Narrow Band Internet of Things Testbed for Proximity Services
Get PDFIn this paper, we present the results of this deployment, indicating that the NB-IoT device successfully connects and communicates with the eNodeB. Session logs show that the EPC effectively initiates the session and authenticates the user equipment (UE). Additionally, the eNodeB establishes a successful connection with the UE based on the parameters defined by the EPC. Wireshark traces demonstrate that the UE can transmit data to the server via an internet connection, with an average latency of 40 ms. Through this work, we explore the benefits of proximity services for NB-IoT networks, providing a valuable platform for experimental testing and prototyping. With the evolution of the Internet of Things (IoT), Narrowband IoT (NB-IoT) has emerged to provide IoT connectivity over existing cellular networks, utilizing limited resources while facing an increased risk of outages at the cell edge. In this work, we developed a testbed for NB-IoT systems to implement the innovative idea of enabling proximity services. This allows devices beyond the coverage area of the eNodeB to transmit their data using a network relay. The testbed is constructed with a software-defined radio functioning as the eNodeB, which wirelessly connects to the NB-IoT node at the front end and to the Evolved Packet Core (EPC) at the back end. The EPC is implemented on a Linux machine using open-source software, while the NB-IoT node is realized with commercially available devices
Exploring Deep Learning Approaches for Early Detection of Chronic Kidney Disease: Trends and Techniques
No full textThis study investigates the application of deep learning models, namely CNN, RNNs, and MLP, for the early prediction of CKD. Early detection of CKD is critical for initiating timely treatment, as the disease can advance with few symptoms. The research leverages a preprocessed Kaggle dataset, divided for training and testing, to assess model performance. Among the models, CNN achieved an impressive 99% accuracy, highlighting its strong feature extraction capabilities. The RNN and MLP models also demonstrated high accuracy, reinforcing the potential of deep learning in enhancing CKD screening processes. This approach can support more personalized and preventive healthcare, potentially improving patient outcomes through earlier interventions.
Keywords: RNN, CKD, Deep Learning, CNN, ANN, LSTM, Performance Optimization
Abbreviation
Full Form
CKD
Chronic Kidney Disease
CNN
Convolutional Neural Network
RNN
Recurrent Neural Network
MLP
Multi-Layer Perceptron
ANN
Artificial Neural Network
LSTM
Long Short-Term Memory (a type of RNN