Search results for: machine translation (MT)

Authors: Derlis Gregor, Kevin Cikel, Mario Arzamendia, Raúl Gregor

Abstract:

This paper presents a self-sustaining mobile system for counting and classification of vehicles through processing video. It proposes a counting and classification algorithm divided in four steps that can be executed multiple times in parallel in a SBC (Single Board Computer), like the Raspberry Pi 2, in such a way that it can be implemented in real time. The first step of the proposed algorithm limits the zone of the image that it will be processed. The second step performs the detection of the mobile objects using a BGS (Background Subtraction) algorithm based on the GMM (Gaussian Mixture Model), as well as a shadow removal algorithm using physical-based features, followed by morphological operations. In the first step the vehicle detection will be performed by using edge detection algorithms and the vehicle following through Kalman filters. The last step of the proposed algorithm registers the vehicle passing and performs their classification according to their areas. An auto-sustainable system is proposed, powered by batteries and photovoltaic solar panels, and the data transmission is done through GPRS (General Packet Radio Service)eliminating the need of using external cable, which will facilitate it deployment and translation to any location where it could operate. The self-sustaining trailer will allow the counting and classification of vehicles in specific zones with difficult access.

Keywords: intelligent transportation system, object detection, vehicle couting, vehicle classification, video processing

Procedia PDF Downloads 319

Authors: Duncan Wallace, M-Tahar Kechadi

Abstract:

In recent years, Machine Learning (ML) approaches have been successfully applied to an analysis of patient symptom data in the context of disease diagnosis, at least where such data is well codified. However, much of the data present in Electronic Health Records (EHR) are unlikely to prove suitable for classic ML approaches. Furthermore, as scores of data are widely spread across both hospitals and individuals, a decentralized, computationally scalable methodology is a priority. The focus of this paper is to develop a method to predict outliers in an out-of-hours healthcare provision center (OOHC). In particular, our research is based upon the early identification of patients who have underlying conditions which will cause them to repeatedly require medical attention. OOHC act as an ad-hoc delivery of triage and treatment, where interactions occur without recourse to a full medical history of the patient in question. Medical histories, relating to patients contacting an OOHC, may reside in several distinct EHR systems in multiple hospitals or surgeries, which are unavailable to the OOHC in question. As such, although a local solution is optimal for this problem, it follows that the data under investigation is incomplete, heterogeneous, and comprised mostly of noisy textual notes compiled during routine OOHC activities. Through the use of Deep Learning methodologies, the aim of this paper is to provide the means to identify patient cases, upon initial contact, which are likely to relate to such outliers. To this end, we compare the performance of Long Short-Term Memory, Gated Recurrent Units, and combinations of both with Convolutional Neural Networks. A further aim of this paper is to elucidate the discovery of such outliers by examining the exact terms which provide a strong indication of positive and negative case entries. While free-text is the principal data extracted from EHRs for classification, EHRs also contain normalized features. Although the specific demographical features treated within our corpus are relatively limited in scope, we examine whether it is beneficial to include such features among the inputs to our neural network, or whether these features are more successfully exploited in conjunction with a different form of a classifier. In this section, we compare the performance of randomly generated regression trees and support vector machines and determine the extent to which our classification program can be improved upon by using either of these machine learning approaches in conjunction with the output of our Recurrent Neural Network application. The output of our neural network is also used to help determine the most significant lexemes present within the corpus for determining high-risk patients. By combining the confidence of our classification program in relation to lexemes within true positive and true negative cases, with an inverse document frequency of the lexemes related to these cases, we can determine what features act as the primary indicators of frequent-attender and non-frequent-attender cases, providing a human interpretable appreciation of how our program classifies cases.

Keywords: artificial neural networks, data-mining, machine learning, medical informatics

Procedia PDF Downloads 128

Authors: John Lamb, Robert Epstein, Vasundhara L. Bhupathi, Sanjeev Kumar Marimekala

Abstract:

As we enter the new era of Artificial Intelligence (AI) and Cloud Computing, we mostly rely on the Machine and Natural Language Processing capabilities of AI, and Energy Efficient Hardware and Software Devices in almost every industry sector. In these industry sectors, much emphasis is on developing new and innovative methods for producing and conserving energy and sustaining the depletion of natural resources. The core pillars of sustainability are economic, environmental, and social, which is also informally referred to as the 3 P's (People, Planet and Profits). The 3 P's play a vital role in creating a core Sustainability Model in the Enterprise. Natural resources are continually being depleted, so there is more focus and growing demand for renewable energy. With this growing demand, there is also a growing concern in many industries on how to reduce carbon emissions and conserve natural resources while adopting sustainability in corporate business models and policies. In our paper, we would like to discuss the driving forces such as Climate changes, Natural Disasters, Pandemic, Disruptive Technologies, Corporate Policies, Scaled Business Models and Emerging social media and AI platforms that influence the 3 main pillars of Sustainability (3P’s). Through this paper, we would like to bring an overall perspective on enterprise strategies and the primary focus on bringing cultural shifts in adapting energy-efficient operational models. Overall, many industries across the globe are incorporating core sustainability principles such as reducing energy costs, reducing greenhouse gas (GHG) emissions, reducing waste and increasing recycling, adopting advanced monitoring and metering infrastructure, reducing server footprint and compute resources (Shared IT services, Cloud computing, and Application Modernization) with the vision for a sustainable environment.

Keywords: climate change, pandemic, disruptive technology, government policies, business model, machine learning and natural language processing, AI, social media platform, cloud computing, advanced monitoring, metering infrastructure

Procedia PDF Downloads 108

Authors: Rezvan Ravanfar Haghighi, V. C. Vani, Suresh Perumal, Sabyasachi Chatterjee, Pratik Kumar

Abstract:

X-ray attenuation coefficient [µ(E)] of any substance, for energy (E), is a sum of the contributions from the Compton scattering [ μCom(E)] and photoelectric effect [µPh(E)]. In terms of the, electron density (ρe) and the effective atomic number (Zeff) we have µCom(E) is proportional to [(ρe)fKN(E)] while µPh(E) is proportional to [(ρeZeffx)/Ey] with fKN(E) being the Klein-Nishina formula, with x and y being the exponents for photoelectric effect. By taking the sample's HU at two different excitation voltages (V=V1, V2) of the CT machine, we can solve for X=ρe, Y=ρeZeffx from these two independent equations, as is attempted in DECT inversion. Since µCom(E) and µPh(E) are both energy dependent, the coefficients of inversion are also dependent on (a) the source spectrum S(E,V) and (b) the detector efficiency D(E) of the CT machine. In the present paper we tabulate these coefficients of inversion for different practical manifestations of S(E,V) and D(E). The HU(V) values from the CT follow: <µ(V)>=<µw(V)>[1+HU(V)/1000] where the subscript 'w' refers to water and the averaging process <….> accounts for the source spectrum S(E,V) and the detector efficiency D(E). Linearity of μ(E) with respect to X and Y implies that (a) <µ(V)> is a linear combination of X and Y and (b) for inversion, X and Y can be written as linear combinations of two independent observations <µ(V1)>, <µ(V2)> with V1≠V2. These coefficients of inversion would naturally depend upon S(E, V) and D(E). We numerically investigate this dependence for some practical cases, by taking V = 100 , 140 kVp, as are used for cardiological investigations. The S(E,V) are generated by using the Boone-Seibert source spectrum, being superposed on aluminium filters of different thickness lAl with 7mm≤lAl≤12mm and the D(E) is considered to be that of a typical Si[Li] solid state and GdOS scintilator detector. In the values of X and Y, found by using the calculated inversion coefficients, errors are below 2% for data with solutions of glycerol, sucrose and glucose. For low Zeff materials like propionic acid, Zeffx is overestimated by 20% with X being within1%. For high Zeffx materials like KOH the value of Zeffx is underestimated by 22% while the error in X is + 15%. These imply that the source may have additional filtering than the aluminium filter specified by the manufacturer. Also it is found that the difference in the values of the inversion coefficients for the two types of detectors is negligible. The type of the detector does not affect on the DECT inversion algorithm to find the unknown chemical characteristic of the scanned materials. The effect of the source should be considered as an important factor to calculate the coefficients of inversion.

Keywords: attenuation coefficient, computed tomography, photoelectric effect, source spectrum

Procedia PDF Downloads 398

Authors: Yanting Cao, Kazumitsu Nawata

Abstract:

As smartphones upgrading, their software and hardware are getting smarter, so the smartphone-based human activity recognition will be described as more refined, complex, and detailed. In this context, we analyzed a set of experimental data obtained by observing and measuring 30 volunteers with six activities of daily living (ADL). Due to the large sample size, especially a 561-feature vector with time and frequency domain variables, cleaning these intractable features and training a proper model becomes extremely challenging. After a series of feature selection and parameters adjustment, a well-performed SVM classifier has been trained.

Keywords: smart sensors, human activity recognition, artificial intelligence, SVM

Procedia PDF Downloads 140

Authors: V. Mantey, N. Findlay, I. Maddox

Abstract:

The quality and quantity of global satellite data have been increasing exponentially in recent years as spaceborne systems become more affordable and the sensors themselves become more sophisticated. This is a valuable resource for many applications, including disaster management and relief. However, while more information can be valuable, the volume of data available is impossible to manually examine. Therefore, the question becomes how to extract as much information as possible from the data with limited manpower. Buildings are a key feature of interest in satellite imagery with applications including telecommunications, population models, and disaster relief. Machine learning tools are fast becoming one of the key resources to solve this problem, and models have been developed to detect buildings in optical satellite imagery. However, by and large, most models focus on affluent regions where buildings are generally larger and constructed further apart. This work is focused on the more difficult problem of detection in populated regions. The primary challenge with detecting small buildings in densely populated regions is both the spatial and spectral resolution of the optical sensor. Densely packed buildings with similar construction materials will be difficult to separate due to a similarity in color and because the physical separation between structures is either non-existent or smaller than the spatial resolution. This study finds that training models until they are overfitting the input sample can perform better in these areas than a more robust, generalized model. An overfitted model takes less time to fine-tune from a generalized pre-trained model and requires fewer input data. The model developed for this study has also been fine-tuned using existing, open-source, building vector datasets. This is particularly valuable in the context of disaster relief, where information is required in a very short time span. Leveraging existing datasets means that little to no manpower or time is required to collect data in the region of interest. The training period itself is also shorter for smaller datasets. Requiring less data means that only a few quality areas are necessary, and so any weaknesses or underpopulated regions in the data can be skipped over in favor of areas with higher quality vectors. In this study, a landcover classification model was developed in conjunction with the building detection tool to provide a secondary source to quality check the detected buildings. This has greatly reduced the false positive rate. The proposed methodologies have been implemented and integrated into a configurable production environment and have been employed for a number of large-scale commercial projects, including continent-wide DEM production, where the extracted building footprints are being used to enhance digital elevation models. Overfitted machine learning models are often considered too specific to have any predictive capacity. However, this study demonstrates that, in cases where input data is scarce, overfitted models can be judiciously applied to solve time-sensitive problems.

Keywords: building detection, disaster relief, mask-RCNN, satellite mapping

Procedia PDF Downloads 167

Authors: Lawrence Koech, Ray Everson, Hein Neomagus, Hilary Rutto

Abstract:

Wet Flue gas desulphurization (FGD) systems are commonly used to remove sulphur dioxide from flue gas by contacting it with limestone in aqueous phase which is obtained by dissolution. Dissolution is important as it affects the overall performance of a wet FGD system. In the present study, effects of pH, stirring speed, solid to liquid ratio and acid concentration on the dissolution of limestone using an organic acid (adipic acid) were investigated. This was investigated using the pH stat apparatus. Calcium ions were analyzed at the end of each experiment using Atomic Absorption (AAS) machine.

Keywords: desulphurization, limestone, dissolution, pH stat apparatus

Procedia PDF Downloads 456

Authors: Xabier Olaz Moratinos

Abstract:

Drone navigation is critical, particularly during the initial phases, such as the initial ascension, where pilots may fail due to strong external interferences that could potentially lead to a crash. In this ongoing work, a drone has been successfully trained to perform an ascent of up to 6 meters at speeds with external disturbances pushing it up to 24 mph, with the DQN algorithm managing external forces affecting the system. It has been demonstrated that the system can control its height, position, and stability in all three axes (roll, pitch, and yaw) throughout the process. The learning process is carried out in the Gazebo simulator, which emulates interferences, while ROS is used to communicate with the agent.

Keywords: machine learning, DQN, gazebo, navigation

Procedia PDF Downloads 108

Authors: Dong Seop Lee, Byung Sik Kim

Abstract:

In the 4th Industrial Revolution, various intelligent technologies have been developed in many fields. These artificial intelligence technologies are applied in various services, including disaster management. Disaster information management does not just support disaster work, but it is also the foundation of smart disaster management. Furthermore, it gets historical disaster information using artificial intelligence technology. Disaster information is one of important elements of entire disaster cycle. Disaster information management refers to the act of managing and processing electronic data about disaster cycle from its’ occurrence to progress, response, and plan. However, information about status control, response, recovery from natural and social disaster events, etc. is mainly managed in the structured and unstructured form of reports. Those exist as handouts or hard-copies of reports. Such unstructured form of data is often lost or destroyed due to inefficient management. It is necessary to manage unstructured data for disaster information. In this paper, the Optical Character Recognition approach is used to convert handout, hard-copies, images or reports, which is printed or generated by scanners, etc. into electronic documents. Following that, the converted disaster data is organized into the disaster code system as disaster information. Those data are stored in the disaster database system. Gathering and creating disaster information based on Optical Character Recognition for unstructured data is important element as realm of the smart disaster management. In this paper, Korean characters were improved to over 90% character recognition rate by using upgraded OCR. In the case of character recognition, the recognition rate depends on the fonts, size, and special symbols of character. We improved it through the machine learning algorithm. These converted structured data is managed in a standardized disaster information form connected with the disaster code system. The disaster code system is covered that the structured information is stored and retrieve on entire disaster cycle such as historical disaster progress, damages, response, and recovery. The expected effect of this research will be able to apply it to smart disaster management and decision making by combining artificial intelligence technologies and historical big data.

Keywords: disaster information management, unstructured data, optical character recognition, machine learning

Procedia PDF Downloads 122

Authors: Wenhao Lan, Ning Li, Qiang Tong

Abstract:

To improve the registration accuracy of a source point cloud and template point cloud when the initial relative deflection angle is too large, a PointNetLK algorithm combined with an oriented bounding box (PointNetLK-OBB) is proposed. In this algorithm, the OBB of a 3D point cloud is used to represent the macro feature of source and template point clouds. Under the guidance of the iterative closest point algorithm, the OBB of the source and template point clouds is aligned, and a mirror symmetry effect is produced between them. According to the fitting degree of the source and template point clouds, the mirror symmetry plane is detected, and the optimal rotation and translation of the source point cloud is obtained to complete the 3D point cloud registration task. To verify the effectiveness of the proposed algorithm, a comparative experiment was performed using the publicly available ModelNet40 dataset. The experimental results demonstrate that, compared with PointNetLK, PointNetLK-OBB improves the registration accuracy of the source and template point clouds when the initial relative deflection angle is too large, and the sensitivity of the initial relative position between the source point cloud and template point cloud is reduced. The primary contribution of this paper is the use of PointNetLK to avoid the non-convex problem of traditional point cloud registration and leveraging the regularity of the OBB to avoid the local optimization problem in the PointNetLK context.

Keywords: mirror symmetry, oriented bounding box, point cloud registration, PointNetLK-OBB

Procedia PDF Downloads 147

Authors: Mpho Mokoatle, Darlington Mapiye, James Mashiyane, Stephanie Muller, Gciniwe Dlamini

Abstract:

Great advances in high-throughput sequencing technologies have resulted in availability of huge amounts of sequencing data in public and private repositories, enabling a holistic understanding of complex biological phenomena. Sequence data are used for a wide range of applications such as gene annotations, expression studies, personalized treatment and precision medicine. However, this rapid growth in sequence data poses a great challenge which calls for novel data processing and analytic methods, as well as huge computing resources. In this work, a machine and statistical learning approach for DNA sequence classification based on $k$-mer representation of sequence data is proposed. The approach is tested using whole genome sequences of Mycobacterium tuberculosis (MTB) isolates to (i) reduce the size of genomic sequence data, (ii) identify an optimum size of k-mers and utilize it to build classification models, (iii) predict the phenotype from whole genome sequence data of a given bacterial isolate, and (iv) demonstrate computing challenges associated with the analysis of whole genome sequence data in producing interpretable and explainable insights. The classification models were trained on 104 whole genome sequences of MTB isoloates. Cluster analysis showed that k-mers maybe used to discriminate phenotypes and the discrimination becomes more concise as the size of k-mers increase. The best performing classification model had a k-mer size of 10 (longest k-mer) an accuracy, recall, precision, specificity, and Matthews Correlation coeffient of 72.0%, 80.5%, 80.5%, 63.6%, and 0.4 respectively. This study provides a comprehensive approach for resampling whole genome sequencing data, objectively selecting a k-mer size, and performing classification for phenotype prediction. The analysis also highlights the importance of increasing the k-mer size to produce more biological explainable results, which brings to the fore the interplay that exists amongst accuracy, computing resources and explainability of classification results. However, the analysis provides a new way to elucidate genetic information from genomic data, and identify phenotype relationships which are important especially in explaining complex biological mechanisms.

Keywords: AWD-LSTM, bootstrapping, k-mers, next generation sequencing

Procedia PDF Downloads 165

Authors: Darlington Mapiye, Mpho Mokoatle, James Mashiyane, Stephanie Muller, Gciniwe Dlamini

Abstract:

Great advances in high-throughput sequencing technologies have resulted in availability of huge amounts of sequencing data in public and private repositories, enabling a holistic understanding of complex biological phenomena. Sequence data are used for a wide range of applications such as gene annotations, expression studies, personalized treatment and precision medicine. However, this rapid growth in sequence data poses a great challenge which calls for novel data processing and analytic methods, as well as huge computing resources. In this work, a machine and statistical learning approach for DNA sequence classification based on k-mer representation of sequence data is proposed. The approach is tested using whole genome sequences of Mycobacterium tuberculosis (MTB) isolates to (i) reduce the size of genomic sequence data, (ii) identify an optimum size of k-mers and utilize it to build classification models, (iii) predict the phenotype from whole genome sequence data of a given bacterial isolate, and (iv) demonstrate computing challenges associated with the analysis of whole genome sequence data in producing interpretable and explainable insights. The classification models were trained on 104 whole genome sequences of MTB isoloates. Cluster analysis showed that k-mers maybe used to discriminate phenotypes and the discrimination becomes more concise as the size of k-mers increase. The best performing classification model had a k-mer size of 10 (longest k-mer) an accuracy, recall, precision, specificity, and Matthews Correlation coeffient of 72.0 %, 80.5 %, 80.5 %, 63.6 %, and 0.4 respectively. This study provides a comprehensive approach for resampling whole genome sequencing data, objectively selecting a k-mer size, and performing classification for phenotype prediction. The analysis also highlights the importance of increasing the k-mer size to produce more biological explainable results, which brings to the fore the interplay that exists amongst accuracy, computing resources and explainability of classification results. However, the analysis provides a new way to elucidate genetic information from genomic data, and identify phenotype relationships which are important especially in explaining complex biological mechanisms

Keywords: AWD-LSTM, bootstrapping, k-mers, next generation sequencing

Procedia PDF Downloads 156

Authors: Walid A. M. Ghoneim, Hamdy A. Ashour, Asmaa E. Abdo

Abstract:

In this paper, linear analysis of a Switched Reluctance Generator (SRG) model is applied on the most common configurations (4/2, 6/4 and 8/6) for both conventional short-pitched and fully-pitched designs, in order to determine the optimum stator/rotor pole angles at which the maximum output voltage is generated per unit excitation current. This study is focused on SRG analysis and design as a proposed solution for renewable energy applications, such as wind energy conversion systems. The world’s potential to develop the renewable energy technologies through dedicated scientific researches was the motive behind this study due to its positive impact on economy and environment. In addition, the problem of rare earth metals (Permanent magnet) caused by mining limitations, banned export by top producers and environment restrictions leads to the unavailability of materials used for rotating machines manufacturing. This challenge gave authors the opportunity to study, analyze and determine the optimum design of the SRG that has the benefit to be free from permanent magnets, rotor windings, with flexible control system and compatible with any application that requires variable-speed operation. In addition, SRG has been proved to be very efficient and reliable in both low-speed or high-speed applications. Linear analysis was performed using MATLAB simulations based on the (Modified generalized matrix approach) of Switched Reluctance Machine (SRM). About 90 different pole angles combinations and excitation patterns were simulated through this study, and the optimum output results for each case were recorded and presented in detail. This procedure has been proved to be applicable for any SRG configuration, dimension and excitation pattern. The delivered results of this study provide evidence for using the 4-phase 8/6 fully pitched SRG as the main optimum configuration for the same machine dimensions at the same angular speed.

Keywords: generalized matrix approach, linear analysis, renewable applications, switched reluctance generator

Procedia PDF Downloads 193

Authors: Manpreet Kaur, Cheol-Hong Min

Abstract:

The increasing advancement in the robotics equipped with machine vision sensors applied to precision agriculture is a demanding solution for various problems in the agricultural farms. An important issue related with the machine vision system concerns crop row and weed detection. This paper proposes an automatic furrow detection system based on real-time processing for identifying crop rows in maize fields in the presence of weed. This vision system is designed to be installed on the farming vehicles, that is, submitted to gyros, vibration and other undesired movements. The images are captured under image perspective, being affected by above undesired effects. The goal is to identify crop rows for vehicle navigation which includes weed removal, where weeds are identified as plants outside the crop rows. The images quality is affected by different lighting conditions and gaps along the crop rows due to lack of germination and wrong plantation. The proposed image processing method consists of four different processes. First, image segmentation based on HSV (Hue, Saturation, Value) decision tree. The proposed algorithm used HSV color space to discriminate crops, weeds and soil. The region of interest is defined by filtering each of the HSV channels between maximum and minimum threshold values. Then the noises in the images were eliminated by the means of hybrid median filter. Further, mathematical morphological processes, i.e., erosion to remove smaller objects followed by dilation to gradually enlarge the boundaries of regions of foreground pixels was applied. It enhances the image contrast. To accurately detect the position of crop rows, the region of interest is defined by creating a binary mask. The edge detection and Hough transform were applied to detect lines represented in polar coordinates and furrow directions as accumulations on the angle axis in the Hough space. The experimental results show that the method is effective.

Keywords: furrow detection, morphological, HSV, Hough transform

Procedia PDF Downloads 228

Authors: Abdulmalek Ahmed

Abstract:

Formation pressure is the main function that affects drilling operation economically and efficiently. Knowing the pore pressure and the parameters that affect it will help to reduce the cost of drilling process. Many empirical models reported in the literature were used to calculate the formation pressure based on different parameters. Some of these models used only drilling parameters to estimate pore pressure. Other models predicted the formation pressure based on log data. All of these models required different trends such as normal or abnormal to predict the pore pressure. Few researchers applied artificial intelligence (AI) techniques to predict the formation pressure by only one method or a maximum of two methods of AI. The objective of this research is to predict the pore pressure based on both drilling parameters and log data namely; weight on bit, rotary speed, rate of penetration, mud weight, bulk density, porosity and delta sonic time. A real field data is used to predict the formation pressure using five different artificial intelligence (AI) methods such as; artificial neural networks (ANN), radial basis function (RBF), fuzzy logic (FL), support vector machine (SVM) and functional networks (FN). All AI tools were compared with different empirical models. AI methods estimated the formation pressure by a high accuracy (high correlation coefficient and low average absolute percentage error) and outperformed all previous. The advantage of the new technique is its simplicity, which represented from its estimation of pore pressure without the need of different trends as compared to other models which require a two different trend (normal or abnormal pressure). Moreover, by comparing the AI tools with each other, the results indicate that SVM has the advantage of pore pressure prediction by its fast processing speed and high performance (a high correlation coefficient of 0.997 and a low average absolute percentage error of 0.14%). In the end, a new empirical correlation for formation pressure was developed using ANN method that can estimate pore pressure with a high precision (correlation coefficient of 0.998 and average absolute percentage error of 0.17%).

Keywords: Artificial Intelligence (AI), Formation pressure, Artificial Neural Networks (ANN), Fuzzy Logic (FL), Support Vector Machine (SVM), Functional Networks (FN), Radial Basis Function (RBF)

Procedia PDF Downloads 147

Authors: Luo Jiaoyang, Yu Hongyang

Abstract:

In this paper, we explore a method that can be used in the working scene of intelligent industrial robots to confirm the identity information of operators to ensure that the robot executes instructions in a sufficiently safe environment. This approach uses three information modalities, namely visible light, depth, and sound. We explored a variety of fusion modes for the three modalities and finally used the joint feature learning method to improve the performance of the model in the case of noise compared with the single-modal case, making the maximum noise in the experiment. It can also maintain an accuracy rate of more than 90%.

Keywords: multimodal, kinect, machine learning, distance image

Procedia PDF Downloads 77

Authors: A. Sopharak, B. Uyyanonvara, S. Barman

Abstract:

Diabetic retinopathy is characterized by the development of retinal microaneurysms. The damage can be prevented if disease is treated in its early stages. In this paper, we are comparing Support Vector Machine (SVM) and Naïve Bayes (NB) classifiers for automatic microaneurysm detection in images acquired through non-dilated pupils. The Nearest Neighbor classifier is used as a baseline for comparison. Detected microaneurysms are validated with expert ophthalmologists’ hand-drawn ground-truths. The sensitivity, specificity, precision and accuracy of each method are also compared.

Keywords: diabetic retinopathy, microaneurysm, naive Bayes classifier, SVM classifier

Procedia PDF Downloads 325

Authors: Dolapo Adeyemo

Abstract:

Technology can be designed specifically for geriatrics and persons with disabilities or ICT accessibility solutions. Both solutions stand to benefit from advances in Artificial intelligence, which are computer systems that perform tasks that require human intelligence. Tasks such as decision making, visual perception, speech recognition, and even language translation are useful in both situation and will provide significant benefits to people with temporarily or permanent disabilities. This research’s goal is to review innovations focused on the use of artificial intelligence that bridges the accessibility gap in technology from a user-centered perspective. A mixed method approach that utilized a comprehensive review of academic literature on the subject combined with semi structure interviews of users, developers, and technology product owners. The internet of things and artificial intelligence technology is creating new opportunities in the assistive technology space and proving accessibility to existing technology. Device now more adaptable to the needs of the user by learning the behavior of users as they interact with the internet. Accessibility to devices have witnessed significant enhancements that continue to benefit people with disabilities. Examples of other advances identified are prosthetic limbs like robotic arms supported by artificial intelligence, route planning software for the visually impaired, and decision support tools for people with disabilities and even clinicians that provide care.

Keywords: ICT, IOT, accessibility solutions, universal design

Procedia PDF Downloads 85

Authors: Nidhi Gopal

Abstract:

Routing is the process of moving a packet of data from source to destination and enables messages to pass from one computer to another and eventually reach the target machine. A router is a networking device that forwards data packets between computer networks. It is connected to two or more data lines from different networks (as opposed to a network switch, which connects data lines from one single network). This paper mainly emphasizes upon the study of router device, its top level architecture, and how various sub-modules of router i.e. Register, FIFO, FSM and Synchronizer are synthesized, and simulated and finally connected to its top module.

Keywords: data packets, networking, router, routing

Procedia PDF Downloads 806

Authors: Debasish Pradhan, Shaktiprasad Pradhan, Rakesh Kumar Pradhan, Gitanjali Tripathy

Abstract:

Suppressors of cytokine signalling (SOCS1), a newly indentified antiapoptotic molecule is a downstream effector of the receptor tyrosine kinase-Ras signalling pathway. The current study has uncovered that SOCS1 may have wide and imperative capacities, particularly because of its close correlation with malignant tumors. To investigate the impact of SOCS1 on MDR, we analyzed the expression of P-gp and SOCS1 by immunohistochemistry and found there was a positive correlation between them. At that point, we effectively interfered with RNA translation by the contamination of siRNA of SOCS1 into MCF7/ADM breast cancer cell lines through a lentivirus, and the expression of the target gene was significantly inhibited. After RNAi, the drug resistance was reduced altogether and the expression of MDR1 mRNA and P-gp in MCF7/ADM cell lines demonstrated a significant decrease. Likewise, the expression of P53 protein increased in a statistically significant manner (p ≤ 0.01) after RNAi exposure. Moreover, flow cytometry analysis uncovers that cell cycle and anti-apoptotic enhancing capacity of cells changed after RNAi treatment. These outcomes proposed SOCS1 may take part in breast cancer MDR by managing MDR1 and P53 expression, changing cell cycle and enhancing the anti-apoptotic ability.

Keywords: breast cancer, multidrug resistance, SOCS1 gene, MDR1 gene, RNA interference

Procedia PDF Downloads 354

Authors: Lou J. Pino, Mark Campbell, Matthew J. Kennedy, Ashleigh C. Kennedy

Abstract:

A concussion is complex and nuanced, with cognitive rest being a key component of recovery. Cognitive overexertion during rehabilitation from a concussion is associated with delayed recovery. However, daily living imposes cognitive demands that may be unavoidable and difficult to quantify. Therefore, a portable tool capable of alerting patients before cognitive overexertion occurs could allow patients to maintain their quality of life while preventing symptoms and recovery setbacks. EEG allows for a sensitive measure of cognitive exertion. Clinical 32-lead EEG headsets are not practical for day-to-day concussion rehabilitation management. However, there are now commercially available and affordable portable EEG headsets. Thus, these headsets can potentially be used to continuously monitor cognitive exertion during mental tasks to alert the wearer of overexertion, with the aim of preventing the occurrence of symptoms to speed recovery times. The objective of this study was to test an algorithm for predicting cognitive exertion from EEG data collected from a portable headset. EEG data were acquired from 10 participants (5 males, 5 females). Each participant wore a portable 4 channel EEG headband while completing 10 tasks: rest (eyes closed), rest (eyes open), three levels of the increasing difficulty of logic puzzles, three levels of increasing difficulty in multiplication questions, rest (eyes open), and rest (eyes closed). After each task, the participant was asked to report their perceived level of cognitive exertion using the NASA Task Load Index (TLX). Each participant then completed a second session on a different day. A customized machine learning model was created using data from the first session. The performance of each model was then tested using data from the second session. The mean correlation coefficient between TLX scores and predicted cognitive exertion was 0.75 ± 0.16. The results support the efficacy of the algorithm for predicting cognitive exertion. This demonstrates that the algorithms developed in this study used with portable EEG devices have the potential to aid in the concussion recovery process by monitoring and warning patients of cognitive overexertion. Preventing cognitive overexertion during recovery may reduce the number of symptoms a patient experiences and may help speed the recovery process.

Keywords: cognitive activity, EEG, machine learning, personalized recovery

Procedia PDF Downloads 218

Authors: Farah Al-Fityani, Aljohara Alsowail, Shatha Bindawood, Heba Balrbeah

Abstract:

Locket is a popular app that lets users share spontaneous photos with a close circle of friends. The app offers a unique way to stay connected with loved ones by allowing users to see glimpses of their day through photos displayed on a widget on their home screen. This summary outlines the process of developing an app like Locket, highlighting the importance of user privacy and security. It also details the findings of a study on user engagement with the Locket app, revealing positive sentiment towards its features and concept but also identifying areas for improvement. Overall, the summary portrays Locket as a successful app that is changing the way people connect on social media.

Keywords: locket, app, machine learning, connect

Procedia PDF Downloads 43

Authors: Xiaoli Zhang, Xiongfei Li, Yuncong Feng

Abstract:

ROC curves, as a widely used evaluating tool in machine learning field, are the tradeoff of true positive rate and negative rate. However, they are blamed for ignoring some vital information in the evaluation process, such as the amount of information about the target that each instance carries, predicted score given by each classification model to each instance. Hence, in this paper, a new classification performance method is proposed by extending the Receiver Operator Characteristic (ROC) curves to 3D space, which is denoted as 3D ROC Histogram. In the histogram, the

Keywords: classification, performance evaluation, receiver operating characteristic histogram, hardness prediction

Procedia PDF Downloads 309

Authors: Sisaynew Tesfaw Admassu

Abstract:

The strengthening of timber beams can be necessary for several reasons including the increase of live loads (possible in a historical building for a change of destination of use or upgrading to meet new requirements), the reduction of the resistant cross-sections following deterioration (attacks of biological agents such as fungi, and insects) or traumatic events (fires) and the excess of deflection in the members. The main purpose of strengthening an element is not merely to repair it, but also to prevent and minimize the appearance of future problems. This study did an experimental investigation on the behavior of reference and strengthened solid timber beams. The strengthening materials used in this study were CSM-450 glass fiber and steel materials for both flexural and shear strengthening techniques. Twenty-two solid timber beams of Juniperus procera (TID) species with the dimensions of 60 x 90 x 780 mm were used in the present study. The binding material to bond the strengthening materials with timber was general-purpose resin with Luperox® K10 MEKP catalyst. Three beams were used as control beams (unstrengthen beams) while the remaining nineteen beams were strengthened using the strengthening materials for flexure and shear. All the beams were tested for three points loading to failure by using a Universal Testing Machine, UTM-600kN machine. The experimental results showed that the strengthened beams performed better than the unstrengthen beams. The experimental result of flexural strengthened beams showed that the load-bearing capacity of strengthened beams increased between 16.34 – 42.55%. Four layers of Glass Fiber Reinforced polymer on the tension side of the beams was shown to be the most effective way to enhance load-bearing capacity. The strengthened beams also have an enhancement in their flexural stiffness. The stiffness of flexural strengthened beams was increased between 1.18 – 65.53% as compared to the control beams. The highest increment in stiffness has occurred on beams strengthened using 2x60 mm steel plates. The shear-strengthened beams showed a relatively small amount of performance as compared to flexural-strengthened beams; the reason is that the beams are sufficient for shear. The polyester resin used in the experimental work showed good performance in bonding agents between materials. The resin showed more effectiveness in GFRP materials than steel materials.

Keywords: heritage structures, strengthening, stiffness, adhesive, polyester resin, steel plates

Procedia PDF Downloads 69

Authors: Xabier Olaz Moratinos

Abstract:

Drone navigation is critical, particularly during the initial phases, such as the initial ascension, where pilots may fail due to strong external interferences that could potentially lead to a crash. In this ongoing work, a drone has been successfully trained to perform an ascent of up to 6 meters at speeds with external disturbances pushing it up to 24 mph, with the DQN algorithm managing external forces affecting the system. It has been demonstrated that the system can control its height, position, and stability in all three axes (roll, pitch, and yaw) throughout the process. The learning process is carried out in the Gazebo simulator, which emulates interferences, while ROS is used to communicate with the agent.

Keywords: machine learning, DQN, Gazebo, navigation

Procedia PDF Downloads 73

Authors: Cíntia Martins Sanches

Abstract:

In the first part of Phoenissae of Seneca, Antigone is a guide to Oedipus, and they leave Thebes: he is blind searching for death (inflicting the punishment himself wished on the killer of Laius, ie exile and death); she is trying to convince him to give up such punishment and bring him back to Thebes. Concerning Oedipus lines, we observed a high frequency of Latin neuter in the treatment the protagonist gave to his daughter Antigone. We considered in this study that such frequency may be related to the sanctification of the daughter, who is seen by him as an enlightened being and without defects, free of the human condition (which takes on the existence of failures by essence). This study, thus, puts forward an analysis of the passages the said feature is present, relating them to the effect of meaning found in each occurrence. As part of a doctorate, this study investigates the stylistic idiom of Seneca in the Oedipus and Phoenissae tragedies, aiming at translating both tragedies expressively. The concept of stylistic idiom concerns the stylistic affinity required for a translation to be equivalent to the source text. In this wise, this study inquires into how the Latin text is organized poetically, pointing out the expressive features frequently appearing in both dramas. The method we used is based on the Semiotics theory — observing how connotation, ie a language use in which prevails the poetic function, naturally polysemous, acts to achieve each expressive effect.

Keywords: antigone, neuter, Oedipus, Phoenissae, Seneca

Procedia PDF Downloads 285

Authors: Filip Rusak, Yulia Arzhaeva, Dadong Wang

Abstract:

Machine learning based Computer-Aided Diagnosis (CAD) is gaining much popularity in medical imaging and diagnostic radiology. However, it requires a large amount of high quality and labeled training image datasets. The training images may come from different sources and be acquired from different radiography machines produced by different manufacturers, digital or digitized copies of film radiographs, with various sizes as well as different pixel intensity distributions. In this paper, a content-aware image augmentation method is presented to deal with these variations. The results of the proposed method have been validated graphically by plotting the removed and added seams of pixels on original images. Two different chest X-ray (CXR) datasets are used in the experiments. The CXRs in the datasets defer in size, some are digital CXRs while the others are digitized from analog CXR films. With the proposed content-aware augmentation method, the Seam Carving algorithm is employed to resize CXRs and the corresponding labels in the form of image masks, followed by histogram matching used to normalize the pixel intensities of digital radiography, based on the pixel intensity values of digitized radiographs. We implemented the algorithms, resized the well-known Montgomery dataset, to the size of the most frequently used Japanese Society of Radiological Technology (JSRT) dataset and normalized our digital CXRs for testing. This work resulted in the unified off-the-shelf CXR dataset composed of radiographs included in both, Montgomery and JSRT datasets. The experimental results show that even though the amount of augmentation is large, our algorithm can preserve the important information in lung fields, local structures, and global visual effect adequately. The proposed method can be used to augment training and testing image data sets so that the trained machine learning model can be used to process CXRs from various sources, and it can be potentially used broadly in any medical imaging applications.

Keywords: computer-aided diagnosis, image augmentation, lung segmentation, medical imaging, seam carving

Procedia PDF Downloads 217

Authors: Xiaolin Zeng, Lei Cheng, Zhirong Li, Xueping Liu

Abstract:

Aiming at the problem of low accuracy and susceptibility to cultural relic diseases in the generation of high-resolution archaeology drawings by current image generation algorithms, an archaeology drawings generation algorithm based on a conditional generative adversarial network is proposed. An attention mechanism is added into the high-resolution image generation network as the backbone network, which enhances the line feature extraction capability and improves the accuracy of line drawing generation. A dual-branch parallel architecture consisting of two backbone networks is implemented, where the semantic translation branch extracts semantic features from orthophotographs of cultural relics, and the gradient screening branch extracts effective gradient features. Finally, the fusion fine-tuning module combines these two types of features to achieve the generation of high-quality and high-resolution archaeology drawings. Experimental results on the self-constructed archaeology drawings dataset of grotto temple statues show that the proposed algorithm outperforms current mainstream image generation algorithms in terms of pixel accuracy (PA), structural similarity (SSIM), and peak signal-to-noise ratio (PSNR) and can be used to assist in drawing archaeology drawings.

Keywords: archaeology drawings, digital heritage, image generation, deep learning

Procedia PDF Downloads 51

Authors: Rosa Figueroa, Christopher Flores

Abstract:

Text categorization is the problem of assigning a new document to a set of predetermined categories, on the basis of a training set of free-text data that contains documents whose category membership is known. To train a classification model, it is necessary to extract characteristics in the form of tokens that facilitate the learning and classification process. In text categorization, the feature extraction process involves the use of word sequences also known as N-grams. In general, it is expected that documents belonging to the same category share similar features. The Smith-Waterman (SW) algorithm is a dynamic programming algorithm that performs a local sequence alignment in order to determine similar regions between two strings or protein sequences. This work explores the use of SW algorithm as an alternative to feature extraction in text categorization. The dataset used for this purpose, contains 2,610 annotated documents with the classes Obese/Non-Obese. This dataset was represented in a matrix form using the Bag of Word approach. The score selected to represent the occurrence of the tokens in each document was the term frequency-inverse document frequency (TF-IDF). In order to extract features for classification, four experiments were conducted: the first experiment used SW to extract features, the second one used unigrams (single word), the third one used bigrams (two word sequence) and the last experiment used a combination of unigrams and bigrams to extract features for classification. To test the effectiveness of the extracted feature set for the four experiments, a Support Vector Machine (SVM) classifier was tuned using 20% of the dataset. The remaining 80% of the dataset together with 5-Fold Cross Validation were used to evaluate and compare the performance of the four experiments of feature extraction. Results from the tuning process suggest that SW performs better than the N-gram based feature extraction. These results were confirmed by using the remaining 80% of the dataset, where SW performed the best (accuracy = 97.10%, weighted average F-measure = 97.07%). The second best was obtained by the combination of unigrams-bigrams (accuracy = 96.04, weighted average F-measure = 95.97) closely followed by the bigrams (accuracy = 94.56%, weighted average F-measure = 94.46%) and finally unigrams (accuracy = 92.96%, weighted average F-measure = 92.90%).

Keywords: comorbidities, machine learning, obesity, Smith-Waterman algorithm

Procedia PDF Downloads 294

Authors: Zafer Yüce, Paşa Yayla, Alev Taşkın

Abstract:

There are heaps of analytical methods to estimate the crack growth life of a component. Soft computing methods have an increasing trend in predicting fatigue life. Their ability to build complex relationships and capability to handle huge amounts of data are motivating researchers and industry professionals to employ them for challenging problems. This study focuses on soft computing methods, especially random forest regressors and artificial neural networks with hyperparameter optimization algorithms such as grid search and random grid search, to estimate the crack growth life of an aircraft wing splice joint under variable amplitude loading. TensorFlow and Scikit-learn libraries of Python are used to build the machine learning models for this study. The material considered in this work is 7050-T7451 aluminum, which is commonly preferred as a structural element in the aerospace industry, and regarding the crack type; corner crack is used. A finite element model is built for the joint to calculate fastener loads and stresses on the structure. Since finite element model results are validated with analytical calculations, findings of the finite element model are fed to AFGROW software to calculate analytical crack growth lives. Based on Fighter Aircraft Loading Standard for Fatigue (FALSTAFF), 90 unique fatigue loading spectra are developed for various load levels, and then, these spectrums are utilized as inputs to the artificial neural network and random forest regression models for predicting crack growth life. Finally, the crack growth life predictions of the machine learning models are compared with analytical calculations. According to the findings, a good correlation is observed between analytical and predicted crack growth lives.

Keywords: aircraft, fatigue, joint, life, optimization, prediction.

Procedia PDF Downloads 171