Browse Items (2150 total)
Sort by:
-
A Deep Learning Method for Classification in Brain-Computer Interface
Neural activity is the controlling signal used in enabling BCI to have direct communication with a computer. An array of EEG signals aid in the selection of the neural signal. The feature extractors and classifiers have a specific pattern of EEG control for a given BCI protocol, which is tailor-made and limited to that specific signal. Although a single protocol is applied in the deep neural networks used in EEG-based brain-computer interfaces, which are being used in the feature extraction and classification of speech recognition and computer vision, it is unclear how these architectures find themselves generalized in other area and prototypes. The deep learning approach used in transferring knowledge acquired from the source tasks to the target tasks is called transfer learning. Conventional machine learning algorithms have been surpassed by deep neural networks while solving problems concerning the real world. However, the best deep neural networks were identified by considering the knowledge of the problem domain. A significant amount of time and computational resources have to be spent to validate this approach. This work presents a deep learning neural network architecture based on Visual Geometry Group Network (VGGNet), Residual Network (ResNet), and inception network methods. Experimental results show that the proposed method achieves better performance than other methods. 2023 IEEE. -
A Deep Learning Methodology CNN-ADAM for the Prediction of PCOS from Text Report
Text categorization is a popular piece of work in natural language processing (NLP) and machine learning, and Convolutional Neural Networks (CNNs) can be used effectively for this purpose. Although CNNs are traditionally associated with computer vision tasks, they have been adapted and applied successfully to text classification problems. In the proposed study Convolutional Neural Networks (CNNs) with adam optimization algorithm plays a crucial role in detecting PCOS words from sonographic text reports. 2023 IEEE. -
A Design of Agricultural Robotics for the use of Sowing and Planting
Agricultural robots is always getting better to deal with problems like population growth, fast urbanization, fierce competition for high-quality goods, worries about protecting the environment, and a lack of skilled workers. This in-depth study looks at the main uses of farming robotic systems, covering jobs like preparing the land, sowing, planting, treating plants, gathering, estimating yields, and phenotyping. Each robot is judged on how it moves, what it will be used for, whether it has sensors, a robotic arm, or a computer vision program, as well as its development stage and where it came from. The study finds trends, possible problems, and things that stop business growth by looking at these shared traits. It also shows which countries are putting money into studying and developing (R&D) for these products. The study points out four important areas - movement systems as a whole sensor, computer vision computer programs, and communication technologies - that need more research to make smart agriculture better. The results make it clear that spending money on farming robotic systems can pay off in the long run by helping with things like accurate yield estimates and short-term benefits like keeping an eye on the harvest. 2024 IEEE. -
A Deterministic Key-Frame Indexing and Selection for Surveillance Video Summarization
Video data is voluminous and impacts the data storage devices as there are CCTV surveillance videos being created every minute and stored continuously. Due to this increase in data there is a need to create semantic information out of the frames that are being stored. Video Summarization is a process that continuously monitors changes and helps in reducing the number of frames being stored. This work enables summarization to be carried out based on selecting threshold-based system that can select key-frames ideally suit for storage and further analysis. Initially a Global threshold based on Otsus method is carried out for all frames of a surveillance video and based on the set threshold a retrospective comparison is done on each frame based on statistical methods to converge on determining the keyframes. A similarity index is generated based on the iterative comparison of frames based on global and local threshold comparison. The local threshold is indexed based on Analysing Method Patterns to Locate Errors(AMPLE), An-derbergs D(AbD), Cohens Kappa(CK), Tanimoto Similarity(TS), Tversky feature contrast model(TFCM), Pearson coefficient of mean square contingency(Pmsc). The Global threshold is updated each time a keyframe is selected based on the comparison of local and global threshold. The results are compared with five surveillance videos and six methods to identify keyframes Selection Rate is the metric used for calculating the performance. 2019 IEEE. -
A distributed randomization framework for privacy preservation in big data
The privacy preservation is a big challenge for data generated from various sources such as social networking sites, online transaction, weather forecast to name a few. Due to the socialization of the internet and cloud computing pica bytes of unstructured data is generated online with intrinsic values. The inflow of big data and the requirement to move this information throughout an organization has become a new target for hackers. This data is subject to privacy laws and should be protected. The proposed protocol is one step toward the security in case of above circumstances where data is coming from multiple participants and all are concerned about individual privacy and confidentiality. 2014 IEEE. -
A Dual Step Strategy for Retinal Thin Vessel Enhancement/Extraction
Blood vessel extraction from retinal images is a challenging and fundamental step in pathological analysis. Most of the vessel extraction algorithms face difficulty in the extraction of thin vessels. In this paper, a dual step strategy for retinal thin vessel enhancement/extraction is proposed. Since thin vessel pixels have intensities closer to the background non-vessel pixels, the first level enhancement algorithms usually suffers in its accurate extraction. This led to explore a novel idea of eliminating the effects of thick vessel pixels in a reference image, via replacing it with neighboring non-vessel pixels. By applying second level enhancement on the vessel subtracted image, thin vessels are projected and improvement in extraction is attained subjectively as well as objectively. 2019 IEEE. -
A Dynamic Anomaly Detection Approach for Fault Detection on Fire Alarm System Based on Fuzzy-PSO-CNN Approach
Early detection is crucial due to the catastrophic threats to life and property that are involved with fires. Sensory systems used in fire alarms are prone to false alerts and breakdowns, endangering lives and property. Therefore, it is essential to check the functionality of smoke detectors often. Traditional plans for such systems have included periodic maintenance; however, because they don't account for the condition of the fire alarm sensors, they are sometimes carried out not when necessary but rather on a predefined conservative timeframe. They describe a data-driven online anomaly detection of smoke detectors, which analyzes the behavior of these devices over time and looks for aberrant patterns that may imply a failure, to aid in the development of a predictive maintenance approach. The suggested procedure begins with three steps: preprocessing, segmentation, and model training. A pre-processing unit can enhance data quality by compensating for sensor drifts, sample-to-sample volatility, and disturbances (noise). The proposed approach normalizes the data in preparation. The smoke source can be detected by using segmentation to differentiate it from the background. Following segmentation, Fuzzy-PSO-CNN is used to train the models. CNN and PSO, two of the most used alternatives, are both outperformed by the proposed method. 2023 IEEE. -
A Feature Selection Study on the Bot-IoT Dataset Using Ensemble Classification Techniques
IoT is an emerging giant in the field of technol- ogy, taking over traditional systems, providing interconnected- ness, convenience, efficiency, and automation, making our lives unimaginably better. However, security for these IoT systems is challenging, especially due to their interconnectedness, making them vulnerable to various cyber threats. The rising tide of IoT botnets, especially, presents a unique challenge. This has urgently increased the need for Intrusion Detection research. Modern Intrusion Detection approaches often employ Machine Learning for effective results. Feature Selection is extremely important while creating Machine Learning Classification models to avoid overfitting and poor performance. This paper focuses on running a Feature Selection study on the Bot-IoT dataset provided by UNSW to increase the accuracy of a ML model. The paper tests 5 types of Feature Selection methods, from Filter- based, Wrapper-based and Embedded methods, combined with two distinct ensemble classifiers: Random Forest + Adaboost and XGBoost. Each combination is tested with the dataset, and the accuracy is compared to find the most effective and versatile feature selection method that can assist both Stacking and Voting- type Ensemble classifiers. The results show that Karl Pearson can provide the best accuracy when applied to both Ensemble Classifiers. 2024 IEEE. -
A Feminist Perspective on the Food and Gender based Marketing Narrative
Nutrition to the body is a basic element for sustenance and growth biologically, provided through food. This paper aims to understand why there is a difference between foods that are marketed gender-specifically to males and females separately. There have been a lot of participative changes in the household kitchen activities since the birth of the concept. However, certain things have continued to remain the same either as a result of tradition, preference, or systemic societal loop. This paper aims to categorically understand this patterned behaviour behind gender based food marketing and the consequent consumptions so as to find a more sustainable and inclusive approach for food marketing for the firms of this industry. The aim is also to shed light on the impact of such practices on the psychological level of the individual buyer that stems to form a pattern, creating a recurring practice out of habit, over internal choice. The Electrochemical Society -
A finger print recognition using CNN Model
The fundamental goal of this research is to improve the new identification accuracy for fingerprint acknowledgment by contrasting Convolutional Neural Networks (CNN) model frameworks for biometric safety in the cloud with Conventional inception models (TIM). Accuracy was computed and compared using a CNN model and standard Inception Models (N=10). The statistical significance was calculated using SPSS. Average and standard deviation for a 95% confidence interval, 0.05% G-power cutoff. The TIM and Convolutional Neural Networks performed an autonomous T-Test on the samples. CNN is more successful (93%) than TIM (61%). Based on a significant value of 0.048 for the comparison ratio (p0.05), there is a statistically significant difference between the CNN and the TIM transformation. According to the findings, the suggested CNN model is 93% accurate on the dataset, with no rejected samples. 2023 IEEE. -
A Fog-Based Retrieval of Real-Time Data for Health Applications
Fog computing is an emerging technology that offers high-quality cloud services by providing high bandwidth, low latency, and efficient computational power and storage capacity. Although cloud computing is an efficient solution so far to store and retrieve the huge data of IoT devices, it is expected to limit its performance due to low latency and storage capacity. Fog computing addresses these limitations by extending its services to the cloud at the edge of the network. In this paper, we use a fog computing network approach for efficiently retrieving the real-time patient data. The performance of our proposed approach has been compared with the cloud computing approach in terms of retrieval time of real-time data. 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. -
A Framework for Digital Forensics Using Blockchain to Secure Digital Data
Digital forensics (DF) requires evidence integrity and provenance across boundaries of jurisdiction, and blockchain technology is ideal for ensuring that. As part of this paper, we discussed a digital forensic framework designed to help prevent duplication of data and secure digital data. In order to accomplish such forensic capabilities, we provide a block-based forensics framework. Using it, examinations are validated, irreversible, traceable, robust, and demonstrate high levels of confidence among examiners and evidence entities. 2022 IEEE. -
A Framework for Dress Code Monitoring System using Transfer Learning from Pre-Trained YOLOv4 Model
Maintaining a proper dress code in organizations or any environment is very important. It not only imbibes a sense of discipline but also reflects the personality and qualities of people as individuals. To follow this practice, some organizations like educational institutions and a few corporations have made it mandatory for the personnel to maintain proper attire as per their regulations. Manual checks are performed to adhere to the organizations' regulations which becomes tedious and erroneous most of the times. Having an automated system not only saves time but also there is very little scope of mistakes and errors. Taking this into context, the main aim and idea behind the project is to propose a model for detecting the dress code in such workplaces and educational institutions where the attire needs to be regularly monitored. The model detects Business Formals (Blazer, Shirt & Pants) worn by the personnel, for which CNN has been considered, along with YOLOv4, for performing the detection, due to its nature of giving the highest accuracy in comparison to the other object-detection models. Providing the Mean Average Precision of around 81%, it becomes evident that the model performs quite well in performing the detections. 2023 IEEE. -
A framework for smart transportation using Big Data
In the current era of information technology, data driven decision is widely recognized. It leads to involvement of the term 'Big Data'. The use of IOT and ICT deployment is a key player of the smart city project in India. It leads to smart transportation systems with huge amounts of real time data that needs to be communicated, aggregated, interpreted, analyzed and maintained. These technologies enhance the effective usage of smart transportation systems, which is economical and has a high social impact. Social applications like transportation can be benefited by using IOT, ICT and big data analytics to give better prediction. In this paper, we present how big data analytics can be used to build a smart transportation system. Increasing traffic and frequent jams in today's scenario are becoming a routine, citizens are facing various health issues due to the bad transportation systems such as high blood pressure, stress, asthma due to air and noise pollution. In smart transportation mobility can be easily implemented as most of the citizens use smartphones. It can be easily linked to smart traffic signals to achieve the objective of smart transportation. Smart transportation is a key component to attract companies as it leads to better services, business planning, support beneficial environment and social behavior. 2016 IEEE. -
A fuzzy computing software quality model
Expectation of the quality of a software varies from user to user. A fuzzy approach to measure the quality of a software is very appropriate so that it can deal with non-crisp aspects of the various parameters. In the proposed model, ordered intuitionistic fuzzy soft sets (OIFSS) and relative similarity measures of OIFSS are considered in the backdrop of fuzzy multiple criteria decision making (FMCDM) approach. 2019 Author(s). -
A Gated Recurrent Unit Based Continual Normalization Model for Arrythmia Classification Using ECG Signals
In this world, around 31% of the deaths are commonly caused because of cardiovascular diseases. Around 80% of sudden deaths occur due to cardiac arrhythmias and heart diseases. The mortality rate has increased for cardiac disease and therefore early heart disease detection is significant to preclude patients from dying. At the initial phase, the heart disease is detected by analyzing abnormal heartbeats. The existing models failed to select the features before performing the extraction of features. The developed model examined MIT-BIB database to surpass the overfitting issue. Therefore, in the present research work, the Gated Recurrent Unit (GRU) based Continual Normalization (CN) classifier is used to speed up the training to a higher learning rate to enable simpler learning for the standard deviation of the neurons' output. The extracted features were used to classify Electrocardiogram (ECG) signals into 5 important classes named as N, S, V, F & Q which denote the kinds of arrhythmia. The findings revealed that the proposed GRU based Continual Normalization technique obtained an accuracy of 99.41% which is better when compared with the existing researches. 2023 IEEE. -
A GPS-Gradient Mapped Database-Based Fuzzy Energy Management System for a SeriesParallel Hybrid Electric Vehicle
The Energy Management System developed for the hybrid electric vehicle operates using a database with GPS co-ordinates and corresponding altitudes mapped, thereby giving a predictive control to optimize the operation of the seriesparallel hybrid system. The system aims at extracting the maximum potential of the seriesparallel hybrid power train architecture. The mapping of the latitude and longitude obtained from a global positioning system (GPS) to the altitude measured to create a database which generates a predefined driving cycle prior to the actual motion of the vehicle. The created database is then used in a MATLAB/Simulink model to simulate the operation of the seriesparallel hybrid system and implement the Energy Management System. The validated data is then tested in a Raspberry Pi (RPi)-based prototype. The Energy Management System regulates the vehicle dynamics based on the input drive cycle. The fuzzy logic-based control mechanism is implemented in the RPi to optimize the load sharing between the IC engine and the brushless DC motor. 2020, Springer Nature Singapore Pte Ltd. -
A Heuristic Approach to Resolve Priority-Driven Unbalanced Transportation Problem (PUTP)
This research addresses the priority-driven unbalanced transportation Problem (PUTP), characterized by a situation where the overall demand surpasses the available supply. We propose the Max-flow Min-cost Priority-driven Unbalanced Transportation Problem (MMPUTP) as a heuristic approach to handle this issue effectively. The strategy of MMPUTP focuses on optimizing resource allocation and reducing costs, making it highly effective in fulfilling high priority needs in a cost-efficient manner. Through a comparison with Vogel's Approximation Method (VAM) over different sets of problems ranging in size from 5?5 to 50?50, the effectiveness of the MMPUTP algorithm is evident. The findings underscore the significance of choosing the right algorithm based on the size and complexity of the problem set in the context of the Priority-driven Unbalanced Transportation Problem, with MMPUTP proving to be a flexible and reliable option in various situations. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
A Heuristic Model For Personalised Risk Assesment of PCOS
According to WHO 8-13% of women are affected by Polycystic Ovary Syndrome (PCOS) out of which 70% women remain undiagnosed, it is a common endocrine disorder necessitating early diagnosis for timely intervention. In this paper a heuristic model is developed for PCOS prediction, by combining XGBoost and Random Forest through stacking techniques. Class imbalance was addressed using Random Oversampling. Cross-validation demonstrated the meta-model's superior accuracy compared to individual XGBoost and Random Forest models, highlighting its potential for reliable PCOS prediction. It is observed that the best possible results that the meta-model was able to provide was a score of 93.5% which was acquired in the 4th sample, the lowest score was 87.90% attained in the 2nd sample. To finalise the results, the mean accuracy was calculated which is 90.98% with a standard deviation of 1.96. deterministic model offers reproducible results and interpretability, aiding clinical decision-making. Future research could explore additional biomarkers and probabilistic techniques for personalized risk assessment. 2024 IEEE. -
A Hybrid Approach for Predictive Maintenance Monitoring of Aircraft Engines
The realm of aircraft maintenance involves predictive maintenance, which utilizes historical data and machine parts' performance to anticipate the need for maintenance activities. The primary focus of this paper is to delve into the application of predictive maintenance of aircraft gas turbine engines. Our methodology involves assigning a randomly chosen deterioration value and monitoring the change in flow and efficiency over time. By carefully analyzing these factors, we can deduce whether the engines are at fault and whether their condition will deteriorate further. The ultimate objective is to identify potential engine malfunctions early to prevent future accidents. Recent years have witnessed the emergence of multiple machine learning and deep learning algorithms to predict the Remaining Useful Life (RUL) of engines. The precision and accuracy of these algorithms in assessing the performance of aircraft engines are pretty promising. We have incorporated a hybrid model on various time series cycles to enhance their efficacy further. Employing data collected from 21 sensors, we can predict the remaining useful life of the turbine engines with greater precision and accuracy. 2024 IEEE.