Search results for: fault detection and classification

Authors: Sofia Ait Bouazza

Abstract:

We consider the problem of finding a minimum identifying code in a graph. This problem was initially introduced in 1998 and has been since fundamentally connected to a wide range of applications (fault diagnosis, location detection …). Suppose we have a building into which we need to place fire alarms. Suppose each alarm is designed so that it can detect any fire that starts either in the room in which it is located or in any room that shares a doorway with the room. We want to detect any fire that may occur or use the alarms which are sounding to not only to not only detect any fire but be able to tell exactly where the fire is located in the building. For reasons of cost, we want to use as few alarms as necessary. The first problem involves finding a minimum domination set of a graph. If the alarms are three state alarms capable of distinguishing between a fire in the same room as the alarm and a fire in an adjacent room, we are trying to find a minimum locating domination set. If the alarms are two state alarms that can only sound if there is a fire somewhere nearby, we are looking for a differentiating domination set of a graph. These three areas are the subject of much active research; we primarily focus on the third problem. An identifying code of a graph G is a dominating set C such that every vertex x of G is distinguished from other vertices by the set of vertices in C that are at distance at most r≥1 from x. When only vertices out of the code are asked to be identified, we get the related concept of a locating dominating set. The problem of finding an identifying code (resp a locating dominating code) of minimum size is a NP-hard problem, even when the input graph belongs to a number of specific graph classes. Therefore, we study this problem in some restricted classes of undirected graphs like split graph, line graph and path in a directed graph. Then we present some results on the identifying code by giving an exact value of upper total locating domination and a total 2-identifying code in directed and undirected graph. Moreover we determine exact values of locating dominating code and edge identifying code of thin headless spider and locating dominating code of complete suns.

Keywords: identiying codes, locating dominating set, split graphs, thin headless spider

Procedia PDF Downloads 480

Authors: Sachi Nandan Mohanty, Shreya Sinha, Sweeti Sah, Shweta Sharma

Abstract:

The advancement of natural language processing has majorly influenced the field of medical transcript classification, providing a robust framework for enhancing the accuracy of clinical data processing. It has enormous potential to transform healthcare and improve people's livelihoods. This research focuses on improving the accuracy of medical transcript categorization using Bidirectional Encoder Representations from Transformers (BERT) and its specialized variants, including BioBERT, ClinicalBERT, SciBERT, and BlueBERT. The experimental work employs Optuna, an optimization framework, for hyperparameter tuning to identify the most effective variant, concluding that BioBERT yields the best performance. Furthermore, various optimizers, including Adam, RMSprop, and Layerwise adaptive large batch optimization (LAMB), were evaluated alongside BERT's default AdamW optimizer. The findings show that the LAMB optimizer achieves equally good performance as AdamW. Synthetic data generation techniques from Gretel were utilized to augment the dataset, expanding the original dataset from 5,000 to 10,000 rows. Subsequent evaluations demonstrated that the model maintained its performance with synthetic data, with the LAMB optimizer showing marginally better results. The enhanced dataset and optimized model configurations improved classification accuracy, showcasing the efficacy of the BioBERT variant and the LAMB optimizer. It resulted in an accuracy of up to 98.2% and 90.8% for the original and combined datasets, respectively.

Keywords: BioBERT, clinical data, healthcare AI, transformer models

Procedia PDF Downloads 0

Authors: Evelyn Asante Yeboah, Kwabena O. Asubonteng, Justice Camillus Mensah, Christine Furst

Abstract:

Smallholder-dominated mosaic landscapes in rural Africa are relevant for food production, biodiversity conservation, and climate regulation. Land-use transitions threaten the multifunctionality of such landscapes, especially the production capacity of arable lands resulting in food security challenges. Using land-cover maps derived from maximum likelihood classification of Landsat satellite images for the years 2002, 2015, and 2020, post-classification change detection, landscape metrics, and key informant interviews, the study assessed the implications of rubber plantation expansion and oil business development on the food production capacity of Ahanta West District, Ghana. The analysis reveals that settlement and rubber areas expanded by 5.82% and 10.33% of the landscape area, respectively, between 2002 and 2020. This increase translates into over twice their initial sizes (144% in settlement change and 101% in rubber change). Rubber plantation spread dominates the north and southwestern areas, whereas settlement is widespread in the eastern parts of the landscape. Rubber and settlement expanded at the expense of cropland, palm, and shrublands. Land-use transitions between cropland, palm, and shrubland were targeting each other, but the net loss in shrubland was higher (-17.27%). Isolation, subdivision, connectedness, and patch adjacency indices showed patch consolidation in the landscape configuration from 2002 to 2015 and patch fragmentation from 2015 to 2020. The study also found patches with consistent increasing connectivity in settlement areas indicating the influence of oil discovery developments and fragmentation tendencies in rubber, shrubland, cropland, and palm, indicating springing up of smaller rubber farms, the disappearance of shrubland, and splitting up of cropland and palm areas respectively. The results revealed a trend in land-use transitions in favor of smallholder rubber plantation expansion and oil discovery developments, which suggest serious implications on food production systems and poses a risk for food security and landscape multifunctional characteristics. To ensure sustainability in land uses, this paper recommends the enforcement of legislative instruments governing spatial planning and land use in Ghana as embedded in the 2016 land-use and spatial planning act.

Keywords: food production systems, food security, Ghana’s west coast, land-use transitions, multifunctional rural landscapes

Procedia PDF Downloads 145

Authors: Cherifi Abdelhamid

Abstract:

In optical communication systems, OFDM has received increased attention as a means to overcome various limitations of optical transmission systems such as modal dispersion, relative intensity noise, chromatic dispersion, polarization mode dispersion and self-phase modulation. The multipath dispersion limits the maximum transmission data rates. In this paper we investigate OFDM system where multipath induced intersymbol interference (ISI) is reduced and we increase the number of users by combining OFDM system with OCDMA system using direct detection Incorporate OOC (orthogonal optical code) for minimize a bit error rate.

Keywords: OFDM, OCDMA, OOC (orthogonal optical code), (ISI), prim codes (Pc)

Procedia PDF Downloads 652

Authors: Islam Khalil, Amr El-Kadi

Abstract:

With the increased use of web technologies, microservices, and Application Programming Interface (API) for integration between systems, and with the development of containerization of services on the operating system level as a method of isolating system execution and for easing the deployment and scaling of systems, there is a growing need as well as opportunities for providing platforms that improve the security of such services. In our work, we propose an architecture for a containerization platform that utilizes various concepts derived from the human immune system. The goal of the proposed containerization platform is to introduce the concept of slowing down or throttling suspected malicious digital pathogens (intrusions) to reduce their damage footprint while providing more opportunities for forensic inspection of suspected pathogens in addition to the ability to snapshot, rollback, and recover from possible damage. The proposed platform also leverages existing intrusion detection algorithms by integrating and orchestrating their cooperative operation for more effective intrusion detection. We show how this model reduces the damage footprint of intrusions and gives a greater time window for forensic investigation. Moreover, during our experiments, our proposed platform was able to uncover unintentional system design flaws that resulted in internal DDoS-like attacks by submodules of the system itself rather than external intrusions.

Keywords: containers, human immunity, intrusion detection, security, web services

Procedia PDF Downloads 95

Authors: Nadjah Chergui, Narhimene Boustia

Abstract:

Intrusion Detection Systems are an essential tool for network security infrastructure. However, IDSs have a serious problem which is the generating of massive number of alerts, most of them are false positive ones which can hide true alerts and make the analyst confused to analyze the right alerts for report the true attacks. The purpose behind this paper is to present a formalism model to perform correlation engine by the reduction of false positive alerts basing on vulnerability contextual information. For that, we propose a formalism model based on non-monotonic JClassicδє description logic augmented with a default (δ) and an exception (є) operator that allows a dynamic inference according to contextual information.

Keywords: context, default, exception, vulnerability

Procedia PDF Downloads 259

Authors: Fatima Mokeddem

Abstract:

The phonocardiogram signal (PCG) is a physiological signal that reflects heart mechanical activity, is a promising tool for curious researchers in this field because it is full of indications and useful information for medical diagnosis. PCG segmentation is a basic step to benefit from this signal. Therefore, this paper presents an algorithm that serves the separation of heart sounds and heart murmurs in case they exist in order to use them in several applications and heart sounds analysis. The separation process presents here is founded on three essential steps filtering, envelope detection, and heart sounds segmentation. The algorithm separates the PCG signal into S1 and S2 and extract cardiac murmurs.

Keywords: phonocardiogram signal, filtering, Envelope, Detection, murmurs, heart sounds

Procedia PDF Downloads 141

Authors: Mohamad H. Atyeh, Ahmad Khaldi

Abstract:

The main objective of this research is to measure the relationship between the Kuwait stock Exchange (KSE) index and its two marketing sectors after the new market classification. The findings of this research are important for Public economic policy makers as they need to know if the new system (new classification) is efficient and to what level, to monitor the markets and intervene with appropriate measures. The data used are the daily index of the whole Kuwaiti market and the daily closing price, number of deals and volume of shares traded of two marketing sectors (consumer goods and consumer services) for the period from the 13th of May 2012 till the 12th of December 2016. The results indicate a positive direct impact of the closing price, volume and deals indexes of the consumer goods and the consumer services companies on the overall KSE index, volume and deals of the Kuwaiti stock market (KSE).

Keywords: correlation, market capitalization, Kuwait Stock Exchange (KSE), marketing sectors, stock performance

Procedia PDF Downloads 326

Authors: Swarnadip Ghosh

Abstract:

Sentiment Analysis is the process of detecting the contextual polarity of text. In other words, it determines whether a piece of writing is positive, negative or neutral.Sentiment analysis of documents holds great importance in today's world, when numerous information is stored in databases and in the world wide web. An efficient algorithm to illicit such information, would be beneficial for social, economic as well as medical purposes. In this project, we have developed an algorithm to classify a document into positive or negative. Using our algorithm, we obtained a feature set from the data, and classified the documents based on this feature set. It is important to note that, in the classification, we have not used the independence assumption, which is considered by many procedures like the Naive Bayes. This makes the algorithm more general in scope. Moreover, because of the sparsity and high dimensionality of such data, we did not use empirical distribution for estimation, but developed a method by finding degree of close clustering of the data points. We have applied our algorithm on a movie review data set obtained from IMDb and obtained satisfactory results.

Keywords: sentiment, Run's Test, cross validation, higher dimensional pmf estimation

Procedia PDF Downloads 402

Authors: Malarvizhi Selvaraj, Paul Fergus, Andy Shaw

Abstract:

Uterine contractions produced during labour have the potential to damage the foetus by diminishing the maternal blood flow to the placenta. In order to observe this phenomenon labour and delivery are routinely monitored using cardiotocography monitors. An obstetrician usually makes the diagnosis of foetus hypoxia by interpreting cardiotocography recordings. However, cardiotocography capture and interpretation is time-consuming and subjective, often lead to misclassification that causes damage to the foetus and unnecessary caesarean section. Both of these have a high impact on the foetus and the cost to the national healthcare services. Automatic detection of foetal heart rate may be an objective solution to help to reduce unnecessary medical interventions, as reported in several studies. This paper aim is to provide a system for better identification and interpretation of abnormalities of the fetal heart rate using RStudio. An open dataset of 552 Intrapartum recordings has been filtered with 0.034 Hz filters in an attempt to remove noise while keeping as much of the discriminative data as possible. Features were chosen following an extensive literature review, which concluded with FIGO features such as acceleration, deceleration, mean, variance and standard derivation. The five features were extracted from 552 recordings. Using these features, recordings will be classified either normal or abnormal. If the recording is abnormal, it has got more chances of hypoxia.

Keywords: cardiotocography, foetus, intrapartum, hypoxia

Procedia PDF Downloads 216

Authors: Mohamed Alimoussa, Alice Porebski, Nicolas Vandenbroucke, Rachid Oulad Haj Thami, Sana El Fkihi

Abstract:

Color and texture are highly discriminant visual cues that provide an essential information in many types of images. Color texture representation and classification is therefore one of the most challenging problems in computer vision and image processing applications. Color textures can be represented in different color spaces by using multiple image descriptors which generate a high dimensional set of texture features. In order to reduce the dimensionality of the feature set, feature selection techniques can be used. The goal of feature selection is to find a relevant subset from an original feature space that can improve the accuracy and efficiency of a classification algorithm. Traditionally, feature selection is focused on removing irrelevant features, neglecting the possible redundancy between relevant ones. This is why some feature selection approaches prefer to use feature clustering analysis to aid and guide the search. These techniques can be divided into two categories. i) Feature clustering-based ranking algorithm uses feature clustering as an analysis that comes before feature ranking. Indeed, after dividing the feature set into groups, these approaches perform a feature ranking in order to select the most discriminant feature of each group. ii) Feature clustering-based subset search algorithms can use feature clustering following one of three strategies; as an initial step that comes before the search, binded and combined with the search or as the search alternative and replacement. In this paper, we propose a new feature clustering-based sequential selection approach for the purpose of color texture representation and classification. Our approach is a three step algorithm. First, irrelevant features are removed from the feature set thanks to a class-correlation measure. Then, introducing a new automatic feature clustering algorithm, the feature set is divided into several feature clusters. Finally, a sequential search algorithm, based on a filter model and a separability measure, builds a relevant and non redundant feature subset: at each step, a feature is selected and features of the same cluster are removed and thus not considered thereafter. This allows to significantly speed up the selection process since large number of redundant features are eliminated at each step. The proposed algorithm uses the clustering algorithm binded and combined with the search. Experiments using a combination of two well known texture descriptors, namely Haralick features extracted from Reduced Size Chromatic Co-occurence Matrices (RSCCMs) and features extracted from Local Binary patterns (LBP) image histograms, on five color texture data sets, Outex, NewBarktex, Parquet, Stex and USPtex demonstrate the efficiency of our method compared to seven of the state of the art methods in terms of accuracy and computation time.

Keywords: feature selection, color texture classification, feature clustering, color LBP, chromatic cooccurrence matrix

Procedia PDF Downloads 137

Authors: Johannes Mertl

Abstract:

Our surrounding air contains various particles. Besides typical representatives of inorganic dust, such as soot and ash, also particles originating from animals, microorganisms or plants are floating through the air, so-called bioaerosols. The group of bioaerosols consists of a broad spectrum of particles of different size, including fungi, bacteria, viruses, spores, or tree, flower and grass pollen that are of high relevance for allergy sufferers. In dependence of the environmental climate and the actual season, these allergenic particles can be found in enormous numbers in the air and are inhaled by humans via the respiration tract, with a potential for inflammatory diseases of the airways, such as asthma or allergic rhinitis. As a consequence air filter systems of ventilation and air conditioning devices are required to meet very high standards to prevent, or at least lower the number of allergens and airborne germs entering the indoor air. Still, filter systems are merely classified for their separation rates using well-defined mineral test dust, while no appropriate sufficiently standardized test methods for bioaerosols exist. However, determined separation rates for mineral test particles of a certain size cannot simply be transferred to bioaerosols, as separation efficiency of particularly fine and respirable particles (< 10 microns) is dependent not only on their shape and particle diameter, but also defined by their density and physicochemical properties. For this reason, the OFI developed a test method, which directly enables a testing of filters and filter media for their separation rates on bioaerosols, as well as a classification of filters. Besides allergens from an intact or fractured tree or grass pollen, allergenic proteins bound to particulates, as well as allergenic fungal spores (e.g. Cladosporium cladosporioides), or bacteria can be used to classify filters regarding their separation rates. Allergens passing through the filter can then be detected by highly sensitive immunological assays (ELISA) or in the case of fungal spores by microbiological methods, which allow for the detection of even one single spore passing the filter. The test procedure, which is carried out in laboratory scale, was furthermore validated regarding its sufficiency to cover real life situations by upscaling using air conditioning devices showing great conformity in terms of separation rates. Additionally, a clinical study with allergy sufferers was performed to verify analytical results. Several different air conditioning filters from the car industry have been tested, showing significant differences in their separation rates.

Keywords: airborne germs, allergens, classification of filters, fine dust

Procedia PDF Downloads 253

Authors: Farina Riaz, Shahab Abdulla, Srinjoy Ganguly, Hajime Suzuki, Ravinesh C. Deo, Susan Hopkins

Abstract:

Recently, quantum machine learning has received significant attention. For various types of data, including text and images, numerous quantum machine learning (QML) models have been created and are being tested. Images are exceedingly complex data components that demand more processing power. Despite being mature, classical machine learning still has difficulties with big data applications. Furthermore, quantum technology has revolutionized how machine learning is thought of, by employing quantum features to address optimization issues. Since quantum hardware is currently extremely noisy, it is not practicable to run machine learning algorithms on it without risking the production of inaccurate results. To discover the advantages of quantum versus classical approaches, this research has concentrated on colored image data. Deep learning classification models are currently being created on Quantum platforms, but they are still in a very early stage. Black and white benchmark image datasets like MNIST and Fashion MINIST have been used in recent research. MNIST and CIFAR-10 were compared for binary classification, but the comparison showed that MNIST performed more accurately than colored CIFAR-10. This research will evaluate the performance of the QML algorithm on the colored benchmark dataset CIFAR-10 to advance QML's real-time applicability. However, deep learning classification models have not been developed to compare colored images like Quantum Convolutional Neural Network (QCNN) to determine how much it is better to classical. Only a few models, such as quantum variational circuits, take colored images. The methodology adopted in this research is a hybrid approach by using penny lane as a simulator. To process the 10 classes of CIFAR-10, the image data has been translated into grey scale and the 28 × 28-pixel image containing 10,000 test and 50,000 training images were used. The objective of this work is to determine how much the quantum approach can outperform a classical approach for a comprehensive dataset of color images. After pre-processing 50,000 images from a classical computer, the QCNN model adopted a hybrid method and encoded the images into a quantum simulator for feature extraction using quantum gate rotations. The measurements were carried out on the classical computer after the rotations were applied. According to the results, we note that the QCNN approach is ~12% more effective than the traditional classical CNN approaches and it is possible that applying data augmentation may increase the accuracy. This study has demonstrated that quantum machine and deep learning models can be relatively superior to the classical machine learning approaches in terms of their processing speed and accuracy when used to perform classification on colored classes.

Keywords: CIFAR-10, quantum convolutional neural networks, quantum deep learning, quantum machine learning

Procedia PDF Downloads 129

Authors: Eunsong Lee, Gae Baik Kim, Young Pil Kim

Abstract:

Bisphenol A (BPA) is one of the endocrine disruptors (EDCs), which have been suspected to be associated with reproductive dysfunction and physiological abnormality in human. Since the BPA has been widely used to make plastics and epoxy resins, the leach of BPA from the lining of plastic products has been of major concern, due to its environmental or human exposure issues. The simple detection of BPA based on the self-assembly of aptamer-mediated gold nanoparticles (AuNPs) has been reported elsewhere, yet the detection sensitivity still remains challenging. Here we demonstrate an improved AuNP-based sensor of BPA by using fluorescence-combined AuNP colorimetry in order to overcome the drawback of traditional AuNP sensors. While the anti-BPA aptamer (full length or truncated ssDNA) triggered the self-assembly of unmodified AuNP (citrate-stabilized AuNP) in the presence of BPA at high salt concentrations, no fluorescence signal was observed by the subsequent addition of SYBR Green, due to a small amount of free anti-BPA aptamer. In contrast, the absence of BPA did not cause the self-assembly of AuNPs (no color change by salt-bridged surface stabilization) and high fluorescence signal by SYBP Green, which was due to a large amount of free anti-BPA aptamer. As a result, the quantitative analysis of BPA was achieved using the combination of absorption of AuNP with fluorescence intensity of SYBR green as a function of BPA concentration, which represented more improved detection sensitivity (as low as 1 ppb) than did in the AuNP colorimetric analysis. This method also enabled to detect high BPA in water-soluble extracts from thermal papers with high specificity against BPS and BPF. We suggest that this approach will be alternative for traditional AuNP colorimetric assays in the field of aptamer-based molecular diagnosis.

Keywords: bisphenol A, colorimetric, fluoroscence, gold-aptamer nanobiosensor

Procedia PDF Downloads 188

Authors: Susan Diamond

Abstract:

Machine Learning workloads have traditionally been run in high-performance computing (HPC) environments, where users log in to dedicated machines and utilize the attached GPUs to run training jobs on huge datasets. Training of large neural network models is very resource intensive, and even after exploiting parallelism and accelerators such as GPUs, a single training job can still take days. Consequently, the cost of hardware is a barrier to entry. Even when upfront cost is not a concern, the lead time to set up such an HPC environment takes months from acquiring hardware to set up the hardware with the right set of firmware, software installed and configured. Furthermore, scalability is hard to achieve in a rigid traditional lab environment. Therefore, it is slow to react to the dynamic change in the artificial intelligent industry. Watson Deep Learning as a service, a cloud-based deep learning platform that mitigates the long lead time and high upfront investment in hardware. It enables robust and scalable sharing of resources among the teams in an organization. It is designed for on-demand cloud environments. Providing a similar user experience in a multi-tenant cloud environment comes with its own unique challenges regarding fault tolerance, performance, and security. Watson Deep Learning as a service tackles these challenges and present a deep learning stack for the cloud environments in a secure, scalable and fault-tolerant manner. It supports a wide range of deep-learning frameworks such as Tensorflow, PyTorch, Caffe, Torch, Theano, and MXNet etc. These frameworks reduce the effort and skillset required to design, train, and use deep learning models. Deep Learning as a service is used at IBM by AI researchers in areas including machine translation, computer vision, and healthcare. 

Keywords: deep learning, machine learning, cognitive computing, model training

Procedia PDF Downloads 209

Authors: Suzanne Zazycki, Bamidele Osamika, Heather Craska, Kaelyn Conaway, Reena Murphy, Stephanie Spence

Abstract:

Invasive Plant Species (IPS) are an important component of effective preservation and conservation of natural lands management. IPS are non-native plants which can aggressively encroach upon native species and pose a significant threat to the ecology, public health, and social welfare of a community. The presence of IPS in U.S. nature preserves has caused economic costs, which has estimated to exceed $26 billion a year. While different methods have been identified to control IPS, few methods have been recognized for early detection of IPS. This study examined identified methods for early detection of IPS in Hueston Woods State Nature Preserve. Mixed methods research design was adopted in this four-phased study. The first phase entailed data gathering, the phase described the characteristics and qualities of IPS and the importance of early detection (ED). The second phase explored ED methods, Geographic Information Systems (GIS) and Citizen Science were discovered as ED methods for IPS. The third phase of the study involved the creation of hotspot maps to identify likely areas for IPS growth. While the fourth phase involved testing and evaluating mobile applications that can support the efforts of citizen scientists in IPS detection. Literature reviews were conducted on IPS and ED methods, and four regional experts from ODNR and Miami University were interviewed. A questionnaire was used to gather information about ED methods used across the state. The findings revealed that geospatial methods, including Unmanned Aerial Vehicles (UAVs), Multispectral Satellites (MSS), and Normalized Difference Vegetation Index (NDVI), are not feasible for early detection of IPS, as they require GIS expertise, are still an emerging technology, and are not suitable for every habitat for the ED of IPS. Therefore, Other ED methods options were explored, which include predicting areas where IPS will grow, which can be done through monitoring areas that are like the species’ native habitat. Through literature review and interviews, IPS are known to grow in frequently disturbed areas such as along trails, shorelines, and streambanks. The research team called these areas “hotspots” and created maps of these hotspots specifically for HW NP to support and narrow the efforts of citizen scientists and staff in the ED of IPS. The results further showed that utilizing citizen scientists in the ED of IPS is feasible, especially through single day events or passive monitoring challenges. The study concluded that the creation of hotspot maps to direct the efforts of citizen scientists are effective for the early detection of IPS. Several recommendations were made, among which is the creation of hotspot maps to narrow the ED efforts as citizen scientists continues to work in the preserves and utilize citizen science volunteers to identify and record emerging IPS.

Keywords: early detection, hueston woods state nature preserve, invasive plant species, hotspots

Procedia PDF Downloads 103

Authors: Soha A. Bahanshal, Byung G. Kim

Abstract:

Identification of patients at high risk for hospital readmission is of crucial importance for quality health care and cost reduction. Predicting hospital readmissions among diabetic patients has been of great interest to many researchers and health decision makers. We build a prediction model to predict hospital readmission for diabetic patients within 30 days of discharge. The core of the prediction model is a modified k Nearest Neighbor called Hybrid Fuzzy Weighted k Nearest Neighbor algorithm. The prediction is performed on a patient dataset which consists of more than 70,000 patients with 50 attributes. We applied data preprocessing using different techniques in order to handle data imbalance and to fuzzify the data to suit the prediction algorithm. The model so far achieved classification accuracy of 80% compared to other models that only use k Nearest Neighbor.

Keywords: machine learning, prediction, classification, hybrid fuzzy weighted k-nearest neighbor, diabetic hospital readmission

Procedia PDF Downloads 186

Authors: Erick C. Valverde, Wansu Lim

Abstract:

Facial emotion recognition (FER) system has been recently developed for more advanced computer vision applications. The ability to identify human emotions would enable smart healthcare facility to diagnose mental health illnesses (e.g., depression and stress) as well as better human social interactions with smart technologies. The FER system involves two steps: 1) face detection task and 2) facial emotion recognition task. It classifies the human expression in various categories such as angry, disgust, fear, happy, sad, surprise, and neutral. This system requires intensive research to address issues with human diversity, various unique human expressions, and variety of human facial features due to age differences. These issues generally affect the ability of the FER system to detect human emotions with high accuracy. Early stage of FER systems used simple supervised classification task algorithms like K-nearest neighbors (KNN) and artificial neural networks (ANN). These conventional FER systems have issues with low accuracy due to its inefficiency to extract significant features of several human emotions. To increase the accuracy of FER systems, deep learning (DL)-based methods, like convolutional neural networks (CNN), are proposed. These methods can find more complex features in the human face by means of the deeper connections within its architectures. However, the inference speed and computational costs of a DL-based FER system is often disregarded in exchange for higher accuracy results. To cope with this drawback, an optimized DL-based FER system is proposed in this study.An extreme version of Inception V3, known as Xception model, is leveraged by applying different network optimization methods. Specifically, network pruning and quantization are used to enable lower computational costs and reduce memory usage, respectively. To support low resource requirements, a 68-landmark face detector from Dlib is used in the early step of the FER system.Furthermore, a DL compiler is utilized to incorporate advanced optimization techniques to the Xception model to improve the inference speed of the FER system. In comparison to VGG-Net and ResNet50, the proposed optimized DL-based FER system experimentally demonstrates the objectives of the network optimization methods used. As a result, the proposed approach can be used to create an efficient and real-time FER system.

Keywords: deep learning, face detection, facial emotion recognition, network optimization methods

Procedia PDF Downloads 118

Authors: Chandra Prayaga, Lakshmi Prayaga, Aaron Wade, Kyle Rank, Gopi Shankar Mallu, Sri Satya, Harsha Pola

Abstract:

In today's tech-driven world, where connectivity shapes our daily lives, maintaining physical and emotional health is crucial. Athletic trainers play a vital role in optimizing athletes' performance and preventing injuries. However, a shortage of trainers impacts the quality of care. This study introduces a vision-based exercise monitoring system leveraging Google's MediaPipe library for precise tracking of bicep curl exercises and simultaneous posture monitoring. We propose a three-stage methodology: landmark detection, side detection, and angle computation. Our system calculates angles at the elbow, wrist, neck, and torso to assess exercise form. Experimental results demonstrate the system's effectiveness in distinguishing between good and partial repetitions and evaluating body posture during exercises, providing real-time feedback for precise fitness monitoring.

Keywords: physical health, athletic trainers, fitness monitoring, technology driven solutions, Google’s MediaPipe, landmark detection, angle computation, real-time feedback

Procedia PDF Downloads 66

Authors: Endang Lukitaningsih, Fathul Jannah, Arief R. Hakim, Ratna D. Puspita, Zullies Ikawati

Abstract:

Rifampicin is a semisynthetic antibiotic derivative of rifamycin B produced by Streptomyces mediterranei. RIF has been used worldwide as first line drug-prescribed throughout tuberculosis therapy. This study aims to develop and to validate an HPLC method couple with a UV detection for determination of rifampicin in spiked human plasma and its application for bioequivalence study. The chromatographic separation was achieved on an RP-C18 column (LachromHitachi, 250 x 4.6 mm., 5μm), utilizing a mobile phase of phosphate buffer/acetonitrile (55:45, v/v, pH 6.8 ± 0.1) at a flow of 1.5 mL/min. Detection was carried out at 337 nm by using spectrophotometer. The developed method was statistically validated for the linearity, accuracy, limit of detection, limit of quantitation, precise and specifity. The specifity of the method was ascertained by comparing chromatograms of blank plasma and plasma containing rifampicin; the matrix and rifampicin were well separated. The limit of detection and limit of quantification were 0.7 µg/mL and 2.3 µg/mL, respectively. The regression curve of standard was linear (r > 0.999) over a range concentration of 20.0 – 100.0 µg/mL. The mean recovery of the method was 96.68 ± 8.06 %. Both intraday and interday precision data showed reproducibility (R.S.D. 2.98% and 1.13 %, respectively). Therefore, the method can be used for routine analysis of rifampicin in human plasma and in bioequivalence study. The validated method was successfully applied in pharmacokinetic and bioequivalence study of rifampicin tablet in a limited number of subjects (under an Ethical Clearance No. KE/FK/6201/EC/2015). The mean values of Cmax, Tmax, AUC(0-24) and AUC(o-∞) for the test formulation of rifampicin were 5.81 ± 0.88 µg/mL, 1.25 hour, 29.16 ± 4.05 µg/mL. h. and 29.41 ± 4.07 µg/mL. h., respectively. Meanwhile for the reference formulation, the values were 5.04 ± 0.54 µg/mL, 1.31 hour, 27.20 ± 3.98 µg/mL.h. and 27.49 ± 4.01 µg/mL.h. From bioequivalence study, the 90% CIs for the test formulation/reference formulation ratio for the logarithmic transformations of Cmax and AUC(0-24) were 97.96-129.48% and 99.13-120.02%, respectively. According to the bioequivamence test guidelines of the European Commission-European Medicines Agency, it can be concluded that the test formulation of rifampicin is bioequivalence with the reference formulation.

Keywords: validation, HPLC, plasma, bioequivalence

Procedia PDF Downloads 291

Authors: Subhajit Das, Nirjhar Dhang

Abstract:

Structural damage detection is a challenging work in the field of structural health monitoring (SHM). The damage detection methods mainly focused on the determination of the location and severity of the damage. Model updating is a well known method to locate and quantify the damage. In this method, an error function is defined in terms of difference between the signal measured from ‘experiment’ and signal obtained from undamaged finite element model. This error function is minimised with a proper algorithm, and the finite element model is updated accordingly to match the measured response. Thus, the damage location and severity can be identified from the updated model. In this paper, an error function is defined in terms of modal data viz. frequencies and modal assurance criteria (MAC). MAC is derived from Eigen vectors. This error function is minimized by teaching-learning-based optimization (TLBO) algorithm, and the finite element model is updated accordingly to locate and quantify the damage. Damage is introduced in the model by reduction of stiffness of the structural member. The ‘experimental’ data is simulated by the finite element modelling. The error due to experimental measurement is introduced in the synthetic ‘experimental’ data by adding random noise, which follows Gaussian distribution. The efficiency and robustness of this method are explained through three examples e.g., one truss, one beam and one frame problem. The result shows that TLBO algorithm is efficient to detect the damage location as well as the severity of damage using modal data.

Keywords: damage detection, finite element model updating, modal assurance criteria, structural health monitoring, teaching learning based optimization

Procedia PDF Downloads 215

Authors: Amarachukwu B. Isiaka, Vivian N. Anakwenze, Chinyere C. Ezemba, Chiamaka R. Ilodinso, Chikodili G. Anaukwu, Chukwuebuka M. Ezeokoli, Ugonna H. Uzoka

Abstract:

Infectious diseases continue to pose significant threats to global public health, necessitating advanced and timely detection methods for effective outbreak management. This study explores the integration of artificial intelligence (AI) in the early detection and management of infectious disease outbreaks. Leveraging vast datasets from diverse sources, including electronic health records, social media, and environmental monitoring, AI-driven algorithms are employed to analyze patterns and anomalies indicative of potential outbreaks. Machine learning models, trained on historical data and continuously updated with real-time information, contribute to the identification of emerging threats. The implementation of AI extends beyond detection, encompassing predictive analytics for disease spread and severity assessment. Furthermore, the paper discusses the role of AI in predictive modeling, enabling public health officials to anticipate the spread of infectious diseases and allocate resources proactively. Machine learning algorithms can analyze historical data, climatic conditions, and human mobility patterns to predict potential hotspots and optimize intervention strategies. The study evaluates the current landscape of AI applications in infectious disease surveillance and proposes a comprehensive framework for their integration into existing public health infrastructures. The implementation of an AI-driven early detection system requires collaboration between public health agencies, healthcare providers, and technology experts. Ethical considerations, privacy protection, and data security are paramount in developing a framework that balances the benefits of AI with the protection of individual rights. The synergistic collaboration between AI technologies and traditional epidemiological methods is emphasized, highlighting the potential to enhance a nation's ability to detect, respond to, and manage infectious disease outbreaks in a proactive and data-driven manner. The findings of this research underscore the transformative impact of harnessing AI for early detection and management, offering a promising avenue for strengthening the resilience of public health systems in the face of evolving infectious disease challenges. This paper advocates for the integration of artificial intelligence into the existing public health infrastructure for early detection and management of infectious disease outbreaks. The proposed AI-driven system has the potential to revolutionize the way we approach infectious disease surveillance, providing a more proactive and effective response to safeguard public health.

Keywords: artificial intelligence, early detection, disease surveillance, infectious diseases, outbreak management

Procedia PDF Downloads 66

Authors: Serkan Solmaz, Jean-Baptiste Gouriet, Nicolas Van de Wyer, Christophe Schram

Abstract:

Efficiency of the cooling process for cryogenic propellant boiling in engine cooling channels on space applications is relentlessly affected by the phase change occurs during the boiling. The effectiveness of the cooling process strongly pertains to the type of the boiling regime such as nucleate and film. Geometric constraints like a non-transparent cooling channel unable to use any of visualization methods. The ultrasonic (US) technique as a non-destructive method (NDT) has therefore been applied almost in every engineering field for different purposes. Basically, the discontinuities emerge between mediums like boundaries among different phases. The sound wave emitted by the US transducer is both transmitted and reflected through a gas-liquid interface which makes able to detect different phases. Due to the thermal and structural concerns, it is impractical to sustain a direct contact between the US transducer and working fluid. Hence the transducer should be located outside of the cooling channel which results in additional interfaces and creates ambiguities on the applicability of the present method. In this work, an exploratory research is prompted so as to determine detection ability and applicability of the US technique on the cryogenic boiling process for a cooling cycle where the US transducer is taken place outside of the channel. Boiling of the cryogenics is a complex phenomenon which mainly brings several hindrances for experimental protocol because of thermal properties. Thus substitute materials are purposefully selected based on such parameters to simplify experiments. Aside from that, nucleate and film boiling regimes emerging during the boiling process are simply simulated using non-deformable stainless steel balls, air-bubble injection apparatuses and air clearances instead of conducting a real-time boiling process. A versatile detection algorithm is perennially developed concerning exploratory studies afterward. According to the algorithm developed, the phases can be distinguished 99% as no-phase, air-bubble, and air-film presences. The results show the detection ability and applicability of the US technique for an exploratory purpose.

Keywords: Ultrasound, ultrasonic, multiphase flow, boiling, cryogenics, detection algorithm

Procedia PDF Downloads 170

Authors: Amanda T. Badaró, Douglas F. Barbin, Sofia T. Garcia, Maria Teresa P. S. Clerici, Amanda R. Ferreira

Abstract:

Food fortification is the intentional addition of a nutrient in a food matrix and has been widely used to overcome the lack of nutrients in the diet or increasing the nutritional value of food. Fortified food must meet the demand of the population, taking into account their habits and risks that these foods may cause. Wheat and its by-products, such as semolina, has been strongly indicated to be used as a food vehicle since it is widely consumed and used in the production of other foods. These products have been strategically used to add some nutrients, such as fibers. Methods of analysis and quantification of these kinds of components are destructive and require lengthy sample preparation and analysis. Therefore, the industry has searched for faster and less invasive methods, such as Near-Infrared Spectroscopy (NIR). NIR is a rapid and cost-effective method, however, it is based on indirect measurements, yielding high amount of data. Therefore, NIR spectroscopy requires calibration with mathematical and statistical tools (Chemometrics) to extract analytical information from the corresponding spectra, as Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). PCA is well suited for NIR, once it can handle many spectra at a time and be used for non-supervised classification. Advantages of the PCA, which is also a data reduction technique, is that it reduces the data spectra to a smaller number of latent variables for further interpretation. On the other hand, LDA is a supervised method that searches the Canonical Variables (CV) with the maximum separation among different categories. In LDA, the first CV is the direction of maximum ratio between inter and intra-class variances. The present work used a portable infrared spectrometer (NIR) for identification and classification of pure and fiber-fortified semolina samples. The fiber was added to semolina in two different concentrations, and after the spectra acquisition, the data was used for PCA and LDA to identify and discriminate the samples. The results showed that NIR spectroscopy associate to PCA was very effective in identifying pure and fiber-fortified semolina. Additionally, the classification range of the samples using LDA was between 78.3% and 95% for calibration and 75% and 95% for cross-validation. Thus, after the multivariate analysis such as PCA and LDA, it was possible to verify that NIR associated to chemometric methods is able to identify and classify the different samples in a fast and non-destructive way.

Keywords: Chemometrics, fiber, linear discriminant analysis, near-infrared spectroscopy, principal component analysis, semolina

Procedia PDF Downloads 212

Authors: Anum Ali, Kashaf ad Dooja, Asif Saleem

Abstract:

The future smart cities trend will be towards Internet of Things (IoT); IoT creates dynamic connections in a ubiquitous manner. Smart cities offer ease and flexibility for daily life matters. By using small devices that are connected to cloud servers based on IoT, network traffic between these devices is growing exponentially, whose security is a concerned issue, since ratio of cyber attack may make the network traffic vulnerable. This paper discusses the latest machine learning approaches in related work further to tackle the increasing rate of cyber attacks, machine learning algorithm is applied to IoT-based network traffic data. The proposed algorithm train itself on data and identify different sections of devices interaction by using supervised learning which is considered as a classifier related to a specific IoT device class. The simulation results clearly identify the attacks and produce fewer false detections.

Keywords: IoT, traffic security, deep learning, classification

Procedia PDF Downloads 154

Authors: Masoud Fathi Kazerouni, Jens Schlemper, Klaus-Dieter Kuhnert

Abstract:

Green planet is one of the Earth’s names which is known as a terrestrial planet and also can be named the fifth largest planet of the solar system as another scientific interpretation. Plants do not have a constant and steady distribution all around the world, and even plant species’ variations are not the same in one specific region. Presence of plants is not only limited to one field like botany; they exist in different fields such as literature and mythology and they hold useful and inestimable historical records. No one can imagine the world without oxygen which is produced mostly by plants. Their influences become more manifest since no other live species can exist on earth without plants as they form the basic food staples too. Regulation of water cycle and oxygen production are the other roles of plants. The roles affect environment and climate. Plants are the main components of agricultural activities. Many countries benefit from these activities. Therefore, plants have impacts on political and economic situations and future of countries. Due to importance of plants and their roles, study of plants is essential in various fields. Consideration of their different applications leads to focus on details of them too. Automatic recognition of plants is a novel field to contribute other researches and future of studies. Moreover, plants can survive their life in different places and regions by means of adaptations. Therefore, adaptations are their special factors to help them in hard life situations. Weather condition is one of the parameters which affect plants life and their existence in one area. Recognition of plants in different weather conditions is a new window of research in the field. Only natural images are usable to consider weather conditions as new factors. Thus, it will be a generalized and useful system. In order to have a general system, distance from the camera to plants is considered as another factor. The other considered factor is change of light intensity in environment as it changes during the day. Adding these factors leads to a huge challenge to invent an accurate and secure system. Development of an efficient plant recognition system is essential and effective. One important component of plant is leaf which can be used to implement automatic systems for plant recognition without any human interface and interaction. Due to the nature of used images, characteristic investigation of plants is done. Leaves of plants are the first characteristics to select as trusty parts. Four different plant species are specified for the goal to classify them with an accurate system. The current paper is devoted to principal directions of the proposed methods and implemented system, image dataset, and results. The procedure of algorithm and classification is explained in details. First steps, feature detection and description of visual information, are outperformed by using Scale invariant feature transform (SIFT), HARRIS-SIFT, and FAST-SIFT methods. The accuracy of the implemented methods is computed. In addition to comparison, robustness and efficiency of results in different conditions are investigated and explained.

Keywords: SIFT combination, feature extraction, feature detection, natural images, natural plant recognition, HARRIS-SIFT, FAST-SIFT

Procedia PDF Downloads 276

Authors: Ali Ulvi Uzer

Abstract:

Data reduction is an important topic in the field of pattern recognition applications. The basic concept is the reduction of multitudinous amounts of data down to the meaningful parts. The Principal Component Analysis (PCA) method is frequently used for data reduction. The Support Vector Machine (SVM) method is a discriminative classifier formally defined by a separating hyperplane. In other words, given labeled training data, the algorithm outputs an optimal hyperplane which categorizes new examples. This study offers a hybrid approach that uses the PCA for data reduction and Support Vector Machines (SVM) for classification. In order to detect the accuracy of the suggested system, two boreholes taken from the soil sample was used. The classification accuracies for this dataset were obtained through using ten-fold cross-validation method. As the results suggest, this system, which is performed through size reduction, is a feasible system for faster recognition of dataset so our study result appears to be very promising.

Keywords: feature selection, sequential forward selection, support vector machines, soil sample

Procedia PDF Downloads 455

Authors: Tatiana Tunon, Gottfried Pestal

Abstract:

Government agencies responsible for the management of fisheries resources publish many types of grey literature, and these materials are increasingly accessible to the public on agency websites. However, scope and quality vary considerably, and end-users need meta-data about the report series when deciding whether to use the information (e.g. apply the methods, include the results in a systematic review), or when prioritizing materials for archiving (e.g. library holdings, reference databases). A proposed taxonomy for these report series was developed based on a review of 41 report series from 6 government agencies in 4 countries (Canada, New Zealand, Scotland, and United States). Each report series was categorized according to multiple criteria describing peer-review process, content, and purpose. A robust classification tree was then fitted to these descriptions, and the resulting taxonomic groups were used to compare agency output from 4 countries using reports available in their online repositories.

Keywords: classification tree, fisheries, government, grey literature

Procedia PDF Downloads 283

Authors: Muhammmad Riandhy Anindika Yudhy, Harry Patria, Ramadhani Santoso

Abstract:

Loss of containment is the primary hazard that process safety management is concerned within the oil and gas industry. Escalation to more serious consequences all begins with the loss of containment, starting with oil and gas release from leakage or spillage from primary containment resulting in pool fire, jet fire and even explosion when reacted with various ignition sources in the operations. Therefore, the heart of process safety management is avoiding loss of containment and mitigating its impact through the implementation of safeguards. The most effective safeguard for the case is an early detection system to alert Operations to take action prior to a potential case of loss of containment. The detection system value increases when applied to a long surface pipeline that is naturally difficult to monitor at all times and is exposed to multiple causes of loss of containment, from natural corrosion to illegal tapping. Based on prior researches and studies, detecting loss of containment accurately in the surface pipeline is difficult. The trade-off between cost-effectiveness and high accuracy has been the main issue when selecting the traditional detection method. The current best-performing method, Real-Time Transient Model (RTTM), requires analysis of closely positioned pressure, flow and temperature (PVT) points in the pipeline to be accurate. Having multiple adjacent PVT sensors along the pipeline is expensive, hence generally not a viable alternative from an economic standpoint.A conceptual approach to combine mathematical modeling using computational fluid dynamics and a supervised machine learning model has shown promising results to predict leakage in the pipeline. Mathematical modeling is used to generate simulation data where this data is used to train the leak detection and localization models. Mathematical models and simulation software have also been shown to provide comparable results with experimental data with very high levels of accuracy. While the supervised machine learning model requires a large training dataset for the development of accurate models, mathematical modeling has been shown to be able to generate the required datasets to justify the application of data analytics for the development of model-based leak detection systems for petroleum pipelines. This paper presents a review of key leak detection strategies for oil and gas pipelines, with a specific focus on crude oil applications, and presents the opportunities for the use of data analytics tools and mathematical modeling for the development of robust real-time leak detection and localization system for surface pipelines. A case study is also presented.

Keywords: pipeline, leakage, detection, AI

Procedia PDF Downloads 191

Authors: Jay L. Fu

Abstract:

Patients with Alzheimer's disease progressively lose their memory and thinking skills and, eventually, the ability to carry out simple daily tasks. The disease is irreversible, but early detection and treatment can slow down the disease progression. In this research, publicly available MRI data and demographic data from 373 MRI imaging sessions were utilized to build models to predict dementia. Various machine learning models, including logistic regression, k-nearest neighbor, support vector machine, random forest, and neural network, were developed. Data were divided into training and testing sets, where training sets were used to build the predictive model, and testing sets were used to assess the accuracy of prediction. Key risk factors were identified, and various models were compared to come forward with the best prediction model. Among these models, the random forest model appeared to be the best model with an accuracy of 90.34%. MMSE, nWBV, and gender were the three most important contributing factors to the detection of Alzheimer’s. Among all the models used, the percent in which at least 4 of the 5 models shared the same diagnosis for a testing input was 90.42%. These machine learning models allow early detection of Alzheimer’s with good accuracy, which ultimately leads to early treatment of these patients.

Keywords: Alzheimer's disease, clinical diagnosis, magnetic resonance imaging, machine learning prediction

Procedia PDF Downloads 143