Search results for: Accuracy

Authors: A. P. Kulkarni, P. Randive, A. R. Mache

Abstract:

In today-s era of plasma and laser cutting, machines using oxy-acetylene flame are also meritorious due to their simplicity and cost effectiveness. The objective to devise a Computer controlled Oxy-Fuel profile cutting machine arose from the increasing demand for metal cutting with respect to edge quality, circularity and lesser formation of redeposit material. The System has an 8 bit micro controller based embedded system, which assures stipulated time response. A new window based Application software was devised which takes a standard CAD file .DXF as input and converts it into numerical data required for the controller. It uses VB6 as a front end whereas MS-ACCESS and AutoCAD as back end. The system is designed around AT89C51RD2, powerful 8 bit, ISP micro controller from Atmel and is optimized to achieve cost effectiveness and also maintains the required accuracy and reliability for complex shapes. The backbone of the system is a cleverly designed mechanical assembly along with the embedded system resulting in an accuracy of about 10 microns while maintaining perfect linearity in the cut. This results in substantial increase in productivity. The observed results also indicate reduced inter laminar spacing of pearlite with an increase in the hardness of the edge region.

Keywords: Computer-Control, Profile, Oxy-Fuel.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2538

Authors: Seyoung Kim, Jeongmin Kim, Kwang Ryel Ryu

Abstract:

A database that records average traffic speeds measured at five-minute intervals for all the links in the traffic network of a metropolitan city. While learning from this data the models that can predict future traffic speed would be beneficial for the applications such as the car navigation system, building predictive models for every link becomes a nontrivial job if the number of links in a given network is huge. An advantage of adopting k-nearest neighbor (k-NN) as predictive models is that it does not require any explicit model building. Instead, k-NN takes a long time to make a prediction because it needs to search for the k-nearest neighbors in the database at prediction time. In this paper, we investigate how much we can speed up k-NN in making traffic speed predictions by reducing the amount of data to be searched for without a significant sacrifice of prediction accuracy. The rationale behind this is that we had a better look at only the recent data because the traffic patterns not only repeat daily or weekly but also change over time. In our experiments, we build several different k-NN models employing different sets of features which are the current and past traffic speeds of the target link and the neighbor links in its up/down-stream. The performances of these models are compared by measuring the average prediction accuracy and the average time taken to make a prediction using various amounts of data.

Keywords: Big data, k-NN, machine learning, traffic speed prediction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1330

Authors: Megha Gupta, Nupur Prakash

Abstract:

Identification of plant diseases has been performed using machine learning and deep learning models on the datasets containing images of healthy and diseased plant leaves. The current study carries out an evaluation of some of the deep learning models based on convolutional neural network architectures for identification of plant diseases. For this purpose, the publicly available New Plant Diseases Dataset, an augmented version of PlantVillage dataset, available on Kaggle platform, containing 87,900 images has been used. The dataset contained images of 26 diseases of 14 different plants and images of 12 healthy plants. The CNN models selected for the study presented in this paper are AlexNet, ZFNet, VGGNet (four models), GoogLeNet, and ResNet (three models). The selected models are trained using PyTorch, an open-source machine learning library, on Google Colaboratory. A comparative study has been carried out to analyze the high degree of accuracy achieved using these models. The highest test accuracy and F1-score of 99.59% and 0.996, respectively, were achieved by using GoogLeNet with Mini-batch momentum based gradient descent learning algorithm.

Keywords: comparative analysis, convolutional neural networks, deep learning, plant disease identification

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 557

Authors: Yanqin Chen, Chao Jiang, Chongdu Cho

Abstract:

This research presents the three-dimensional mechanical characteristics of a commercial gas diffusion layer by experiment and simulation results. Although the mechanical performance of gas diffusion layers has attracted much attention, its reliability and accuracy are still a major challenge. With the help of simulation analysis methods, it is beneficial to the gas diffusion layer’s extensive commercial development and the overall stress analysis of proton electrolyte membrane fuel cells during its pre-production design period. Therefore, in this paper, a three-dimensional constitutive model of a commercial gas diffusion layer, including its material stiffness matrix parameters, is developed and coded, in the user-defined material model of a commercial finite element method software for simulation. Then, the model is validated by comparing experimental results as well as simulation outcomes. As a result, both the experimental data and simulation results show a good agreement with each other, with high accuracy.

Keywords: Gas diffusion layer, proton electrolyte membrane fuel cell, stiffness matrix, three-dimensional mechanical characteristics, user-defined material model.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 878

Authors: Ruxandra Barbulescu, Daniel Ioan, Gabriela Ciuprina

Abstract:

The saltatory conduction is the way the action potential is transmitted along a myelinated axon. The potential diffuses along the myelinated compartments and it is regenerated in the Ranvier nodes due to the ion channels allowing the flow across the membrane. For an efficient simulation of populations of neurons, it is important to use reduced order models both for myelinated compartments and for Ranvier nodes and to have control over their accuracy and inner parameters. The paper presents a reduced order model of this neural system which allows an efficient simulation method for the saltatory conduction in myelinated axons. This model is obtained by concatenating reduced order linear models of 1D myelinated compartments and nonlinear 0D models of Ranvier nodes. The models for the myelinated compartments are selected from a series of spatially distributed models developed and hierarchized according to their modeling errors. The extracted model described by a nonlinear PDE of hyperbolic type is able to reproduce the saltatory conduction with acceptable accuracy and takes into account the finite propagation speed of potential. Finally, this model is again reduced in order to make it suitable for the inclusion in large-scale neural circuits.

Keywords: Saltatory conduction, action potential, myelinated compartments, nonlinear, Ranvier nodes, reduced order models, POD.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 782

Authors: Rafał Urbaniak, Robert Kłosowiak, Michał Ciałkowski, Jarosław Bartoszewicz

Abstract:

The Chair of Thermal Engineering at Poznan University of Technology has been conducted research works on the possibilities of using carbon nanostructures in energy and mechanics applications for a couple of years. Those studies have provided results in a form of co-operation with foreign research centres, numerous publications and patent applications. Authors of this paper have studied the influence of multi-walled carbon nanostructures on changes in static friction arising when steel surfaces were moved. Tests were made using the original test stand consisting of automatically controlled inclined plane driven by precise stepper motors. Computer program created in the LabView environment was responsible for monitoring of the stand operation, accuracy of measurements and archiving the obtained results. Such a solution enabled to obtain high accuracy and repeatability of all conducted experiments. Tests and analysis of the obtained results allowed us to determine how additional layers of carbon nanostructures influenced on changes of static friction coefficients. At the same time, we analyzed the potential possibilities of applying nanostructures under consideration in mechanics.

Keywords: Carbon nanotubes, static friction, dynamic friction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1818

Authors: D. Subedi, S. Pradhan

Abstract:

Current transformers are an integral part of power system because it provides a proportional safe amount of current for protection and measurement applications. However, when the power system experiences an abnormal situation leading to huge current flow, then this huge current is proportionally injected to the protection and metering circuit. Since the protection and metering equipment’s are designed to withstand only certain amount of current with respect to time, these high currents pose a risk to man and equipment. Therefore, during such instances, the CT saturation characteristics have a huge influence on the safety of both man and equipment and on the reliability of the protection and metering system. This paper shows the effect of burden on the Accuracy Limiting factor/ Instrument security factor of current transformers and the change in saturation characteristics of the CT’s. The response of the CT to varying levels of overcurrent at different connected burden will be captured using the data acquisition software LabVIEW. Analysis is done on the real time data gathered using LabVIEW. Variation of current transformer saturation characteristics with changes in burden will be discussed.

Keywords: Accuracy limiting factor, burden, current transformer, instrument security factor, saturation characteristics.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3540

Authors: Bangbo Hu

Abstract:

This paper discusses applications of a revolutionary information technology, Geographic Information Systems (GIS), in the field of the history of cartography by examples, including assessing accuracy of early maps, establishing a database of places and historical administrative units in history, integrating early maps in GIS or digital images, and analyzing social, political, and economic information related to production of early maps. GIS provides a new mean to evaluate the accuracy of early maps. Four basic steps using GIS for this type of study are discussed. In addition, several historical geographical information systems are introduced. These include China Historical Geographic Information Systems (CHGIS), the United States National Historical Geographic Information System (NHGIS), and the Great Britain Historical Geographical Information System. GIS also provides digital means to display and analyze the spatial information on the early maps or to layer them with modern spatial data. How GIS relational data structure may be used to analyze social, political, and economic information related to production of early maps is also discussed in this paper. Through discussion on these examples, this paper reveals value of GIS applications in this field.

Keywords: Cartography, GIS, history, maps.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3077

Authors: M. Abdul-Rani, K. K. Chong, A. F. M. Hani, Y. B. Yap, A. Jamil

Abstract:

Imaging is defined as the process of obtaining geometric images either two dimensional or three dimensional by scanning or digitizing the existing objects or products. In this research, it applied to retrieve 3D information of the human skin surface in medical application. This research focuses on analyzing and determining volume of leg ulcers using imaging devices. Volume determination is one of the important criteria in clinical assessment of leg ulcer. The volume and size of the leg ulcer wound will give the indication on responding to treatment whether healing or worsening. Different imaging techniques are expected to give different result (and accuracies) in generating data and images. Midpoint projection algorithm was used to reconstruct the cavity to solid model and compute the volume. Misinterpretation of the results can affect the treatment efficacy. The objectives of this paper is to compare the accuracy between two 3D data acquisition method, which is laser triangulation and structured light methods, It was shown that using models with known volume, that structured-light-based 3D technique produces better accuracy compared with laser triangulation data acquisition method for leg ulcer volume determination.

Keywords: Imaging, Laser Triangulation, Structured Light, Volume Determination.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1467

Authors: Saeid Gharechelou, Ryutaro Tateishi

Abstract:

Earthquake is an inevitable catastrophic natural disaster. The damages of buildings and man-made structures, where most of the human activities occur are the major cause of casualties from earthquakes. A comparison of optical and SAR data is presented in the case of Kathmandu valley which was hardly shaken by 2015-Nepal Earthquake. Though many existing researchers have conducted optical data based estimated or suggested combined use of optical and SAR data for improved accuracy, however finding cloud-free optical images when urgently needed are not assured. Therefore, this research is specializd in developing SAR based technique with the target of rapid and accurate geospatial reporting. Should considers that limited time available in post-disaster situation offering quick computation exclusively based on two pairs of pre-seismic and co-seismic single look complex (SLC) images. The InSAR coherence pre-seismic, co-seismic and post-seismic was used to detect the change in damaged area. In addition, the ground truth data from field applied to optical data by random forest classification for detection of damaged area. The ground truth data collected in the field were used to assess the accuracy of supervised classification approach. Though a higher accuracy obtained from the optical data then integration by optical-SAR data. Limitation of cloud-free images when urgently needed for earthquak evevent are and is not assured, thus further research on improving the SAR based damage detection is suggested. Availability of very accurate damage information is expected for channelling the rescue and emergency operations. It is expected that the quick reporting of the post-disaster damage situation quantified by the rapid earthquake assessment should assist in channeling the rescue and emergency operations, and in informing the public about the scale of damage.

Keywords: Sentinel-1A data, Landsat-8, earthquake damage, InSAR, rapid monitoring, 2015-Nepal earthquake.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1007

Authors: Somnath Dey, Debasis Samanta

Abstract:

Iris-based biometric system is gaining its importance in several applications. However, processing of iris biometric is a challenging and time consuming task. Detection of iris part in an eye image poses a number of challenges such as, inferior image quality, occlusion of eyelids and eyelashes etc. Due to these problems it is not possible to achieve 100% accuracy rate in any iris-based biometric authentication systems. Further, iris detection is a computationally intensive task in the overall iris biometric processing. In this paper, we address these two problems and propose a technique to localize iris part efficiently and accurately. We propose scaling and color level transform followed by thresholding, finding pupil boundary points for pupil boundary detection and dilation, thresholding, vertical edge detection and removal of unnecessary edges present in the eye images for iris boundary detection. Scaling reduces the search space significantly and intensity level transform is helpful for image thresholding. Experimental results show that our approach is comparable with the existing approaches. Following our approach it is possible to detect iris part with 95-99% accuracy as substantiated by our experiments on CASIA Ver-3.0, ICE 2005, UBIRIS, Bath and MMU iris image databases.

Keywords: Iris recognition, iris localization, biometrics, image processing.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 3141

Authors: Suresh Shirgave, Prakash Kulkarni, José Borges

Abstract:

The continuous growth in the size of the World Wide Web has resulted in intricate Web sites, demanding enhanced user skills and more sophisticated tools to help the Web user to find the desired information. In order to make Web more user friendly, it is necessary to provide personalized services and recommendations to the Web user. For discovering interesting and frequent navigation patterns from Web server logs many Web usage mining techniques have been applied. The recommendation accuracy of usage based techniques can be improved by integrating Web site content and site structure in the personalization process.

Herein, we propose semantically enriched Web Usage Mining method for Personalization (SWUMP), an extension to solely usage based technique. This approach is a combination of the fields of Web Usage Mining and Semantic Web. In the proposed method, we envisage enriching the undirected graph derived from usage data with rich semantic information extracted from the Web pages and the Web site structure. The experimental results show that the SWUMP generates accurate recommendations and is able to achieve 10-20% better accuracy than the solely usage based model. The SWUMP addresses the new item problem inherent to solely usage based techniques.

Keywords: Prediction, Recommendation, Semantic Web Usage Mining, Web Usage Mining.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2979

Authors: Negar Riazifar, Nigel G. Stocks

Abstract:

This paper proposes strategies in level crossing (LC) sampling and reconstruction that provide high fidelity signal reconstruction for speech signals; these strategies circumvent the problem of exponentially increasing number of samples as the bit-depth is increased and hence are highly efficient. Specifically, the results indicate that the distribution of the intervals between samples is one of the key factors in the quality of signal reconstruction; including samples with short intervals does not improve the accuracy of the signal reconstruction, whilst samples with large intervals lead to numerical instability. The proposed sampling method, termed reduced conventional level crossing (RCLC) sampling, exploits redundancy between samples to improve the efficiency of the sampling without compromising performance. A reconstruction technique is also proposed that enhances the numerical stability through linear interpolation of samples separated by large intervals. Interpolation is demonstrated to improve the accuracy of the signal reconstruction in addition to the numerical stability. We further demonstrate that the RCLC and interpolation methods can give useful levels of signal recovery even if the average sampling rate is less than the Nyquist rate.

Keywords: Level crossing sampling, numerical stability, speech processing, trigonometric polynomial.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 363

Authors: Agus Harjoko, Sri Hartati, Henry Dwiyasa

Abstract:

There have been numerous implementations of security system using biometric, especially for identification and verification cases. An example of pattern used in biometric is the iris pattern in human eye. The iris pattern is considered unique for each person. The use of iris pattern poses problems in encoding the human iris. In this research, an efficient iris recognition method is proposed. In the proposed method the iris segmentation is based on the observation that the pupil has lower intensity than the iris, and the iris has lower intensity than the sclera. By detecting the boundary between the pupil and the iris and the boundary between the iris and the sclera, the iris area can be separated from pupil and sclera. A step is taken to reduce the effect of eyelashes and specular reflection of pupil. Then the four levels Coiflet wavelet transform is applied to the extracted iris image. The modified Hamming distance is employed to measure the similarity between two irises. This research yields the identification success rate of 84.25% for the CASIA version 1.0 database. The method gives an accuracy of 77.78% for the left eyes of MMU 1 database and 86.67% for the right eyes. The time required for the encoding process, from the segmentation until the iris code is generated, is 0.7096 seconds. These results show that the accuracy and speed of the method is better than many other methods.

Keywords: Biometric, iris recognition, wavelet transform.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1859

Authors: Maha Hammami, Olfa Benouda Sioud

Abstract:

This study investigates the reliability of management earnings forecasts with reference to these two ingredients: verifiability and neutrality. Specifically, we examine the biasedness (or accuracy) of management earnings forecasts and company specific characteristics that can be associated with accuracy. Based on sample of 102 IPO prospectuses published for admission on NYSE Euronext Paris from 2002 to 2010, we found that these forecasts are on average optimistic and two of the five test variables, earnings variability and financial leverage are significant in explaining ex post bias. Acknowledging the possibility that the bias is the result of the managers’ forecasting behavior, we then examine whether managers decide to under-predict, over-predict or forecast accurately for self-serving purposes. Explicitly, we examine the role of financial distress, operating performance, ownership by insiders and the economy state in influencing managers’ forecasting preferences. We find that managers of distressed firms seem to over-predict future earnings. We also find that when managers are given more stock options, they tend to under-predict future earnings. Finally, we conclude that the management earnings forecasts are affected by an intentional bias due to managers’ forecasting preferences.

Keywords: Intentional bias, Management earnings forecasts, neutrality, verifiability.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2202

Authors: E. Irankhah, M. Zarif, E. Mazrooei Rad, K. Ghandehari

Abstract:

Alzheimer's prevalence is on the rise, and the disease comes with problems like cessation of treatment, high cost of treatment, and the lack of early detection methods. The pathology of this disease causes the formation of protein deposits in the brain of patients called plaque amyloid. Generally, the diagnosis of this disease is done by performing tests such as a cerebrospinal fluid, CT scan, MRI, and spinal cord fluid testing, or mental testing tests and eye tracing tests. In this paper, we tried to use the Medial Temporal Atrophy (MTA) method and the Leave One Out (LOO) cycle to extract the statistical properties of the three Fz, Pz, and Cz channels of ERP signals for early diagnosis of this disease. In the process of CT scan images, the accuracy of the results is 81% for the healthy person and 88% for the severe patient. After the process of ERP signaling, the accuracy of the results for a healthy person in the delta band in the Cz channel is 81% and in the alpha band the Pz channel is 90%. In the results obtained from the signal processing, the results of the severe patient in the delta band of the Cz channel were 89% and in the alpha band Pz channel 92%.

Keywords: Alzheimer's disease, image and signal processing, medial temporal atrophy, LOO Cycle.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1983

Authors: Seyed Ali Jafari, Mohammad Ghomi Avili, Emad Benhelal

Abstract:

In this study, a mathematical model was proposed and the accuracy of this model was assessed to predict the growth of Pseudomonas aeruginosa and rhamnolipid production under nitrogen limiting (sodium nitrate) fed-batch fermentation. All of the parameters used in this model were achieved individually without using any data from the literature. The overall growth kinetic of the strain was evaluated using a dual-parallel substrate Monod equation which was described by several batch experimental data. Fed-batch data under different glycerol (as the sole carbon source, C/N=10) concentrations and feed flow rates were used to describe the proposed fed-batch model and other parameters. In order to verify the accuracy of the proposed model several verification experiments were performed in a vast range of initial glycerol concentrations. While the results showed an acceptable prediction for rhamnolipid production (less than 10% error), in case of biomass prediction the errors were less than 23%. It was also found that the rhamnolipid production by P. aeruginosa was more sensitive at low glycerol concentrations. Based on the findings of this work, it was concluded that the proposed model could effectively be employed for rhamnolipid production by this strain under fed-batch fermentation on up to 80 g l- 1 glycerol.

Keywords: Fed-batch culture, glycerol, kinetic parameters, modelling, Pseudomonas aeruginosa, rhamnolipid.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2405

Authors: Zuraidasahana Zulkarnain, Mohd Shafry Mohd Rahim, Nor Anita Fairos Ismail, Mohd Azhar M. Arsad

Abstract:

Handwritten signature is accepted widely as a biometric characteristic for personal authentication. The use of appropriate features plays an important role in determining accuracy of signature verification; therefore, this paper presents a feature based on the geometrical concept. To achieve the aim, triangle attributes are exploited to design a new feature since the triangle possesses orientation, angle and transformation that would improve accuracy. The proposed feature uses triangulation geometric set comprising of sides, angles and perimeter of a triangle which is derived from the center of gravity of a signature image. For classification purpose, Euclidean classifier along with Voting-based classifier is used to verify the tendency of forgery signature. This classification process is experimented using triangular geometric feature and selected global features. Based on an experiment that was validated using Grupo de Senales 960 (GPDS-960) signature database, the proposed triangular geometric feature achieves a lower Average Error Rates (AER) value with a percentage of 34% as compared to 43% of the selected global feature. As a conclusion, the proposed triangular geometric feature proves to be a more reliable feature for accurate signature verification.

Keywords: biometrics, euclidean classifier, feature extraction, offline signature verification, VOTING-based classifier

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1912

Authors: M. Gooroochurn, D. Kerr, K. Bouazza-Marouf, M. Vloeberghs

Abstract:

This paper presents the automated methods employed for extracting craniofacial landmarks in white light images as part of a registration framework designed to support three neurosurgical procedures. The intraoperative space is characterised by white light stereo imaging while the preoperative plan is performed on CT scans. The registration aims at aligning these two modalities to provide a calibrated environment to enable image-guided solutions. The neurosurgical procedures can then be carried out by mapping the entry and target points from CT space onto the patient-s space. The registration basis adopted consists of natural landmarks (eye corner and ear tragus). A 5mm accuracy is deemed sufficient for these three procedures and the validity of the selected registration basis in achieving this accuracy has been assessed by simulation studies. The registration protocol is briefly described, followed by a presentation of the automated techniques developed for the extraction of the craniofacial features and results obtained from tests on the AR and FERET databases. Since the three targeted neurosurgical procedures are routinely used for head injury management, the effect of bruised/swollen faces on the automated algorithms is assessed. A user-interactive method is proposed to deal with such unpredictable circumstances.

Keywords: Face Processing, Craniofacial Feature Extraction, Preoperative to Intraoperative Registration, Registration Basis.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1377

Authors: Cheng Zhang, James Marco, Walid Allafi, Truong Q. Dinh, W. D. Widanage

Abstract:

Equivalent circuit models (ECMs) are widely used in battery management systems in electric vehicles and other battery energy storage systems. The battery dynamics and the model parameters vary under different working conditions, such as different temperature and state of charge (SOC) levels, and therefore online parameter identification can improve the modelling accuracy. This paper presents a way of online ECM parameter identification using a continuous time (CT) estimation method. The CT estimation method has several advantages over discrete time (DT) estimation methods for ECM parameter identification due to the widely separated battery dynamic modes and fast sampling. The presented method can be used for online SOC estimation. Test data are collected using a lithium ion cell, and the experimental results show that the presented CT method achieves better modelling accuracy compared with the conventional DT recursive least square method. The effectiveness of the presented method for online SOC estimation is also verified on test data.

Keywords: Equivalent circuit model, continuous time domain estimation, linear integral filter method, parameter and SOC estimation, recursive least square.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1303

Authors: Apoorva Vinod, Archana Mathur, Snehanshu Saha

Abstract:

Though there exists a plethora of Activation Functions (AFs) used in single and multiple hidden layer Neural Networks (NN), their behavior always raised curiosity, whether used in combination or singly. The popular AFs – Sigmoid, ReLU, and Tanh – have performed prominently well for shallow and deep architectures. Most of the time, AFs are used singly in multi-layered NN, and, to the best of our knowledge, their performance is never studied and analyzed deeply when used in combination. In this manuscript, we experiment on multi-layered NN architecture (both on shallow and deep architectures; Convolutional NN and VGG16) and investigate how well the network responds to using two different AFs (Sigmoid-Tanh, Tanh-ReLU, ReLU-Sigmoid) used alternately against a traditional, single (Sigmoid-Sigmoid, Tanh-Tanh, ReLU-ReLU) combination. Our results show that on using two different AFs, the network achieves better accuracy, substantially lower loss, and faster convergence on 4 computer vision (CV) and 15 Non-CV (NCV) datasets. When using different AFs, not only was the accuracy greater by 6-7%, but we also accomplished convergence twice as fast. We present a case study to investigate the probability of networks suffering vanishing and exploding gradients when using two different AFs. Additionally, we theoretically showed that a composition of two or more AFs satisfies Universal Approximation Theorem (UAT).

Keywords: Activation Function, Universal Approximation function, Neural Networks, convergence.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 91

Authors: Linyu Wang, Furui Huo, Jianhong Xiang

Abstract:

The Doppler shift generated by high-speed movement and multipath effects in the channel are the main reasons for the generation of a time-frequency doubly-selective (DS) channel. There is severe inter-carrier interference (ICI) in the DS channel. Channel estimation for an orthogonal frequency division multiplexing (OFDM) system over a DS channel is very difficult. The simultaneous orthogonal matching pursuit (SOMP) algorithm under distributed compressive sensing theory (DCS-SOMP) has been used in channel estimation for OFDM systems over DS channels. However, the reconstruction accuracy of the DCS-SOMP algorithm is not high enough in the low Signal-to-Noise Ratio (SNR) stage. To solve this problem, in this paper, we propose an improved DCS-SOMP algorithm based on the inner product difference comparison operation (DCS-DCSOMP). The reconstruction accuracy is improved by increasing the number of candidate indexes and designing the comparison conditions of inner product difference. We combine the DCS-DCSOMP algorithm with the basis expansion model (BEM) to reduce the complexity of channel estimation. Simulation results show the effectiveness of the proposed algorithm and its advantages over other algorithms.

Keywords: OFDM, doubly selective, channel estimation, compressed sensing

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 257

Authors: Qasim Idrees, Gao Liangtian, Liu Bo, Miao Yiran

Abstract:

In order to improve the efficiency and accuracy of the pressure hull structure, optimization of underwater vehicle based on response surface methodology, a method for optimizing the design of pressure hull structure was studied. To determine the pressure shell of five dimensions as a design variable, the application of thin shell theory and the Chinese Classification Society (CCS) specification was carried on the preliminary design. In order to optimize variables of the feasible region, different methods were studied and implemented such as Opt LHD method (to determine the design test sample points in the feasible domain space), parametric ABAQUS solution for each sample point response, and the two-order polynomial response for the surface model of the limit load of structures. Based on the ultimate load of the structure and the quality of the shell, the two-generation genetic algorithm was used to solve the response surface, and the Pareto optimal solution set was obtained. The final optimization result was 41.68% higher than that of the initial design, and the shell quality was reduced by about 27.26%. The parametric method can ensure the accuracy of the test and improve the efficiency of optimization.

Keywords: Parameterization, response surface, structure optimization, pressure hull.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1099

Authors: Khairul Jauhari, Achmad Widodo, Ismoyo Haryanto

Abstract:

In this article, the radial displacement error correction capability of a high precision spindle grinding caused by unbalance force was investigated. The spindle shaft is considered as a flexible rotor mounted on two sets of angular contact ball bearing. Finite element methods (FEM) have been adopted for obtaining the equation of motion of the spindle. In this paper, firstly, natural frequencies, critical frequencies, and amplitude of the unbalance response caused by residual unbalance are determined in order to investigate the spindle behaviors. Furthermore, an optimization design algorithm is employed to minimize radial displacement of the spindle which considers dimension of the spindle shaft, the dynamic characteristics of the bearings, critical frequencies and amplitude of the unbalance response, and computes optimum spindle diameters and stiffness and damping of the bearings. Numerical simulation results show that by optimizing the spindle diameters, and stiffness and damping in the bearings, radial displacement of the spindle can be reduced. A spindle about 4 μm radial displacement error can be compensated with 2 μm accuracy. This certainly can improve the accuracy of the product of machining.

Keywords: Error correction, High precision grinding, Optimization, Radial displacement, Spindle.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1751

Authors: Jian Shi, Baoguo Yu, Haonan Jia, Meng Liu, Ping Huang

Abstract:

Adapting to the flexibility of war, a large number of guided weapons launch from aircraft. Therefore, the inertial navigation system loaded in the weapon needs to undergo an alignment process in the air. This article proposes the following methods to the problem of inaccurate modeling of the system under large misalignment angles, the accuracy reduction of filtering caused by outliers, and the noise changes in GPS signals: first, considering the large misalignment errors of Strapdown Inertial Navigation System (SINS)/GPS, a more accurate model is made rather than to make a small-angle approximation, and the Unscented Kalman Filter (UKF) algorithms are used to estimate the state; then, taking into account the impact of GPS noise changes on the fine alignment algorithm, the innovation adaptive filtering algorithm is introduced to estimate the GPS’s noise in real-time; at the same time, in order to improve the anti-interference ability of the air fine alignment algorithm, a robust filtering algorithm based on outlier detection is combined with the air fine alignment algorithm to improve the robustness of the algorithm. The algorithm can improve the alignment accuracy and robustness under interference conditions, which is verified by simulation.

Keywords: Air alignment, fine alignment, inertial navigation system, integrated navigation system, UKF.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 469

Authors: A. F. Ali, D. M. Shawky

Abstract:

Discovering new biological knowledge from the highthroughput biological data is a major challenge to bioinformatics today. To address this challenge, we developed a new approach for protein classification. Proteins that are evolutionarily- and thereby functionally- related are said to belong to the same classification. Identifying protein classification is of fundamental importance to document the diversity of the known protein universe. It also provides a means to determine the functional roles of newly discovered protein sequences. Our goal is to predict the functional classification of novel protein sequences based on a set of features extracted from each protein sequence. The proposed technique used datasets extracted from the Structural Classification of Proteins (SCOP) database. A set of spectral domain features based on Fast Fourier Transform (FFT) is used. The proposed classifier uses multilayer back propagation (MLBP) neural network for protein classification. The maximum classification accuracy is about 91% when applying the classifier to the full four levels of the SCOP database. However, it reaches a maximum of 96% when limiting the classification to the family level. The classification results reveal that spectral domain contains information that can be used for classification with high accuracy. In addition, the results emphasize that sequence similarity measures are of great importance especially at the family level.

Keywords: Bioinformatics, Artificial Neural Networks, Protein Sequence Analysis, Feature Extraction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2302

Authors: Hui-Chuan Chu, Min-Ju Liao, Wei-Kai Cheng, William Wei-Jen Tsai, Yuh-Min Chen

Abstract:

Avoiding learning failures in mathematics e-learning environments caused by emotional problems in students with autism has become an important topic for combining of special education with information and communications technology. This study presents an adaptive emotional adjustment model in mathematics e-learning for students with autism, emphasizing the lack of emotional perception in mathematics e-learning systems. In addition, an emotion classification for students with autism was developed by inducing emotions in mathematical learning environments to record changes in the physiological signals and facial expressions of students. Using these methods, 58 emotional features were obtained. These features were then processed using one-way ANOVA and information gain (IG). After reducing the feature dimension, methods of support vector machines (SVM), k-nearest neighbors (KNN), and classification and regression trees (CART) were used to classify four emotional categories: baseline, happy, angry, and anxious. After testing and comparisons, in a situation without feature selection, the accuracy rate of the SVM classification can reach as high as 79.3-%. After using IG to reduce the feature dimension, with only 28 features remaining, SVM still has a classification accuracy of 78.2-%. The results of this research could enhance the effectiveness of eLearning in special education.

Keywords: Emotion classification, Physiological and facial Expression measures, Students with autism, Mathematics e-learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1726

Authors: Rahil Garnavi, Mohammad Aldeen, M. Emre Celebi, Alauddin Bhuiyan, Constantinos Dolianitis, George Varigos

Abstract:

Automatic segmentation of skin lesions is the first step towards the automated analysis of malignant melanoma. Although numerous segmentation methods have been developed, few studies have focused on determining the most effective color space for melanoma application. This paper proposes an automatic segmentation algorithm based on color space analysis and clustering-based histogram thresholding, a process which is able to determine the optimal color channel for detecting the borders in dermoscopy images. The algorithm is tested on a set of 30 high resolution dermoscopy images. A comprehensive evaluation of the results is provided, where borders manually drawn by four dermatologists, are compared to automated borders detected by the proposed algorithm, applying three previously used metrics of accuracy, sensitivity, and specificity and a new metric of similarity. By performing ROC analysis and ranking the metrics, it is demonstrated that the best results are obtained with the X and XoYoR color channels, resulting in an accuracy of approximately 97%. The proposed method is also compared with two state-of-theart skin lesion segmentation methods.

Keywords: Border detection, Color space analysis, Dermoscopy, Histogram thresholding, Melanoma, Segmentation.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2194

Authors: Gang Li, Ghulam Mubashar Hassan, Arcady Dyskin, Cara MacNish

Abstract:

Digital image correlation (DIC) is a contactless fullfield displacement and strain reconstruction technique commonly used in the field of experimental mechanics. Comparing with physical measuring devices, such as strain gauges, which only provide very restricted coverage and are expensive to deploy widely, the DIC technique provides the result with full-field coverage and relative high accuracy using an inexpensive and simple experimental setup. It is very important to study the natural patterns effect on the DIC technique because the preparation of the artificial patterns is time consuming and hectic process. The objective of this research is to study the effect of using images having natural pattern on the performance of DIC. A systematical simulation method is used to build simulated deformed images used in DIC. A parameter (subset size) used in DIC can have an effect on the processing and accuracy of DIC and even cause DIC to failure. Regarding to the picture parameters (correlation coefficient), the higher similarity of two subset can lead the DIC process to fail and make the result more inaccurate. The pictures with good and bad quality for DIC methods have been presented and more importantly, it is a systematic way to evaluate the quality of the picture with natural patterns before they install the measurement devices.

Keywords: Digital image correlation (DIC), Deformation simulation, Natural pattern, Subset size.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2747

Authors: Faith-Michael E. Uzoka, Boluwaji A. Akinnuwesi, Taiwo Amoo, Flora Aladi, Stephen Fashoto, Moses Olaniyan, Joseph Osuji

Abstract:

The overarching aim of this study is to develop a soft-computing system for the differential diagnosis of tropical diseases. These conditions are of concern to health bodies, physicians, and the community at large because of their mortality rates, and difficulties in early diagnosis due to the fact that they present with symptoms that overlap, and thus become ‘confusable’. We report on the first phase of our study, which focuses on the development of a fuzzy cognitive map model for early differential diagnosis of tropical diseases. We used malaria as a case disease to show the effectiveness of the FCM technology as an aid to the medical practitioner in the diagnosis of tropical diseases. Our model takes cognizance of manifested symptoms and other non-clinical factors that could contribute to symptoms manifestations. Our model showed 85% accuracy in diagnosis, as against the physicians’ initial hypothesis, which stood at 55% accuracy. It is expected that the next stage of our study will provide a multi-disease, multi-symptom model that also improves efficiency by utilizing a decision support filter that works on an algorithm, which mimics the physician’s diagnosis process.

Keywords: Medical diagnosis, tropical diseases, fuzzy cognitive map, decision support filters, malaria differential diagnosis.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2040