Search results for: MNIST dataset

Authors: Gaetano Zazzaro, Angelo Martone, Roberto V. Montaquila, Luigi Pavone

Abstract:

Seizure is the main factor that affects the quality of life of epileptic patients. The diagnosis of epilepsy, and hence the identification of epileptogenic zone, is commonly made by using continuous Electroencephalogram (EEG) signal monitoring. Seizure identification on EEG signals is made manually by epileptologists and this process is usually very long and error prone. The aim of this paper is to describe an automated method able to detect seizures in EEG signals, using knowledge discovery in database process and data mining methods and algorithms, which can support physicians during the seizure detection process. Our detection method is based on Artificial Neural Network classifier, trained by applying the multilayer perceptron algorithm, and by using a software application, called Training Builder that has been developed for the massive extraction of features from EEG signals. This tool is able to cover all the data preparation steps ranging from signal processing to data analysis techniques, including the sliding window paradigm, the dimensionality reduction algorithms, information theory, and feature selection measures. The final model shows excellent performances, reaching an accuracy of over 99% during tests on data of a single patient retrieved from a publicly available EEG dataset.

Keywords: Artificial Neural Network, Data Mining, Electroencephalogram, Epilepsy, Feature Extraction, Seizure Detection, Signal Processing.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1314

Authors: Rodolfo Lorbieski, Silvia Modesto Nassar

Abstract:

Many supervised machine learning tasks require decision making across numerous different classes. Multi-class classification has several applications, such as face recognition, text recognition and medical diagnostics. The objective of this article is to analyze an adapted method of Stacking in multi-class problems, which combines ensembles within the ensemble itself. For this purpose, a training similar to Stacking was used, but with three levels, where the final decision-maker (level 2) performs its training by combining outputs from the tree-based pair of meta-classifiers (level 1) from Bayesian families. These are in turn trained by pairs of base classifiers (level 0) of the same family. This strategy seeks to promote diversity among the ensembles forming the meta-classifier level 2. Three performance measures were used: (1) accuracy, (2) area under the ROC curve, and (3) time for three factors: (a) datasets, (b) experiments and (c) levels. To compare the factors, ANOVA three-way test was executed for each performance measure, considering 5 datasets by 25 experiments by 3 levels. A triple interaction between factors was observed only in time. The accuracy and area under the ROC curve presented similar results, showing a double interaction between level and experiment, as well as for the dataset factor. It was concluded that level 2 had an average performance above the other levels and that the proposed method is especially efficient for multi-class problems when compared to binary problems.

Keywords: Stacking, multi-layers, ensemble, multi-class.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1094

Authors: Bruce X. B. Yu, Yan Liu, Keith C. C. Chan

Abstract:

We propose to record Activities of Daily Living (ADLs) of elderly people using a vision-based system so as to provide better assistive and personalization technologies. Current ADL-related research is based on data collected with help from non-elderly subjects in laboratory environments and the activities performed are predetermined for the sole purpose of data collection. To obtain more realistic datasets for the application, we recorded ADLs for the elderly with data collected from real-world environment involving real elderly subjects. Motivated by the need to collect data for more effective research related to elderly care, we chose to collect data in the room of an elderly person. Specifically, we installed Kinect, a vision-based sensor on the ceiling, to capture the activities that the elderly subject performs in the morning every day. Based on the data, we identified 12 morning activities that the elderly person performs daily. To recognize these activities, we created a HARELCARE framework to investigate into the effectiveness of existing Human Activity Recognition (HAR) algorithms and propose the use of a transfer learning algorithm for HAR. We compared the performance, in terms of accuracy, and training progress. Although the collected dataset is relatively small, the proposed algorithm has a good potential to be applied to all daily routine activities for healthcare purposes such as evidence-based diagnosis and treatment.

Keywords: Daily activity recognition, healthcare, IoT sensors, transfer learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 895

Authors: Matías Micheletto, Rodrigo Santos, Sergio F. Ochoa

Abstract:

In developing countries, most roads in rural areas are dirt road. They require frequent maintenance since they are affected by erosive events, such as rain or wind, and the transit of heavy-weight trucks and machinery. Early detection of damages on the road condition is a key aspect, since it allows to reduce the maintenance time and cost, and also the limitations for other vehicles to travel through. Most proposals that help address this problem require the explicit participation of drivers, a permanent internet connection, or important instrumentation in vehicles or roads. These constraints limit the suitability of these proposals when applied into developing regions, like Latin America. This paper proposes an alternative method, based on unattended crowdsensing, to determine the quality of dirt roads in rural areas. This method involves the use of a mobile application that complements the road condition surveys carried out by organizations in charge of the road network maintenance, giving them early warnings about road areas that could be requiring maintenance. Drivers can also take advantage of the early warnings while they move through these roads. The method was evaluated using information from a public dataset. Although they are preliminary, the results indicate the proposal is potentially suitable to provide awareness about dirt roads condition to drivers, transportation authority and road maintenance companies.

Keywords: Dirt roads automatic quality assessment, collaborative system, unattended crowdsensing method, roads quality awareness provision.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 533

Authors: El Sayed A. Sharara, A. Tsuji, K. Terada

Abstract:

Call centers have been expanding and they have influence on activation in various markets increasingly. A call center’s work is known as one of the most demanding and stressful jobs. In this paper, we propose the fatigue detection system in order to detect burnout of call center agents in the case of a neck pain and upper back pain. Our proposed system is based on the computer vision technique combined skin color detection with the Viola-Jones object detector. To recognize the gesture of hand poses caused by stress sign, the YCbCr color space is used to detect the skin color region including face and hand poses around the area related to neck ache and upper back pain. A cascade of clarifiers by Viola-Jones is used for face recognition to extract from the skin color region. The detection of hand poses is given by the evaluation of neck pain and upper back pain by using skin color detection and face recognition method. The system performance is evaluated using two groups of dataset created in the laboratory to simulate call center environment. Our call center agent burnout detection system has been implemented by using a web camera and has been processed by MATLAB. From the experimental results, our system achieved 96.3% for upper back pain detection and 94.2% for neck pain detection.

Keywords: Call center agents, fatigue, skin color detection, face recognition.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1045

Authors: Kiurski S. Jelena, Kecić S. Vesna, Aksentijević M. Snežana

Abstract:

Samples of tap and wastewater were collected in three offset printing facilities in Novi Sad, Serbia. Ten physicochemical parameters were analyzed within all collected samples: pH, conductivity, m - alkalinity, p - alkalinity, acidity, carbonate concentration, hydrogen carbonate concentration, active oxygen content, chloride concentration and total alkali content. All measurements were conducted using the standard analytical and instrumental methods. Comparing the obtained results for tap water and wastewater, a clear quality difference was noticeable, since all physicochemical parameters were significantly higher within wastewater samples. The study also involves the application of simple linear regression analysis on the obtained dataset. By using software package ORIGIN 5 the pH value was mutually correlated with other physicochemical parameters. Based on the obtained values of Pearson coefficient of determination a strong positive correlation between chloride concentration and pH (r = -0.943), as well as between acidity and pH (r = -0.855) was determined. In addition, statistically significant difference was obtained only between acidity and chloride concentration with pH values, since the values of parameter F (247.634 and 182.536) were higher than F_critical (5.59). In this way, results of statistical analysis highlighted the most influential parameter of water contamination in offset printing, in the form of acidity and chloride concentration. The results showed that variable dependence could be represented by the general regression model: y = a₀ + a₁x+ k, which further resulted with matching graphic regressions.

Keywords: Pollution, printing industry, simple linear regression analysis, wastewater.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1674

Authors: Xiaoshu Lu, Päivi Leino-Arjas, Kustaa Piha, Akseli Aittomäki, Peppiina Saastamoinen, Ossi Rahkonen, Eero Lahelma

Abstract:

Sickness absence represents a major economic and social issue. Analysis of sick leave data is a recurrent challenge to analysts because of the complexity of the data structure which is often time dependent, highly skewed and clumped at zero. Ignoring these features to make statistical inference is likely to be inefficient and misguided. Traditional approaches do not address these problems. In this study, we discuss model methodologies in terms of statistical techniques for addressing the difficulties with sick leave data. We also introduce and demonstrate a new method by performing a longitudinal assessment of long-term absenteeism using a large registration dataset as a working example available from the Helsinki Health Study for municipal employees from Finland during the period of 1990-1999. We present a comparative study on model selection and a critical analysis of the temporal trends, the occurrence and degree of long-term sickness absences among municipal employees. The strengths of this working example include the large sample size over a long follow-up period providing strong evidence in supporting of the new model. Our main goal is to propose a way to select an appropriate model and to introduce a new methodology for analysing sickness absence data as well as to demonstrate model applicability to complicated longitudinal data.

Keywords: Sickness absence, longitudinal data, methodologies, mix-distribution model.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2271

Authors: Youngjae Jin, Daeshik Kim

Abstract:

This paper describes a cycle accurate simulation results of weight values learned by an auto-encoder behavior model in terms of pre-route simulation. Given the results we visualized the first layer representations with natural images. Many common deep learning threads have focused on learning high-level abstraction of unlabeled raw data by unsupervised feature learning. However, in the process of handling such a huge amount of data, the learning method’s computation complexity and time limited advanced research. These limitations came from the fact these algorithms were computed by using only single core CPUs. For this reason, parallel-based hardware, FPGAs, was seen as a possible solution to overcome these limitations. We adopted and simulated the ready-made auto-encoder to design a behavior model in VerilogHDL before designing hardware. With the auto-encoder behavior model pre-route simulation, we obtained the cycle accurate results of the parameter of each hidden layer by using MODELSIM. The cycle accurate results are very important factor in designing a parallel-based digital hardware. Finally this paper shows an appropriate operation of behavior model based pre-route simulation. Moreover, we visualized learning latent representations of the first hidden layer with Kyoto natural image dataset.

Keywords: Auto-encoder, Behavior model simulation, Digital hardware design, Pre-route simulation, Unsupervised feature learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2690

Authors: Viacheslav Shkuratskyy, Aminu Bello Usman, Michael O’Dea, Mujeeb Ur Rehman, Saifur Rahman Sabuj

Abstract:

This paper examines relationships between solar activity and earthquakes, it applied machine learning techniques: K-nearest neighbour, support vector regression, random forest regression, and long short-term memory network. Data from the SILSO World Data Center, the NOAA National Center, the GOES satellite, NASA OMNIWeb, and the United States Geological Survey were used for the experiment. The 23rd and 24th solar cycles, daily sunspot number, solar wind velocity, proton density, and proton temperature were all included in the dataset. The study also examined sunspots, solar wind, and solar flares, which all reflect solar activity, and earthquake frequency distribution by magnitude and depth. The findings showed that the long short-term memory network model predicts earthquakes more correctly than the other models applied in the study, and solar activity is more likely to effect earthquakes of lower magnitude and shallow depth than earthquakes of magnitude 5.5 or larger with intermediate depth and deep depth

.

Keywords: K-Nearest Neighbour, Support Vector Regression, Random Forest Regression, Long Short-Term Memory Network, earthquakes, solar activity, sunspot number, solar wind, solar flares.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 203

Authors: Larry Powell, Seth Polsley, Drew Casey, Tracy Hammond

Abstract:

Human activity recognition (HAR) systems have shown positive performance when recognizing repetitive activities like walking, running, and sleeping. Water-based activities are a reasonably new area for activity recognition. However, water-based activity recognition has largely focused on supporting the elite and competitive swimming population, which already has amazing coordination and proper form. Beginner swimmers are not perfect, and activity recognition needs to support the individual motions to help beginners. Activity recognition algorithms are traditionally built around short segments of timed sensor data. Using a time window input can cause performance issues in the machine learning model. The window’s size can be too small or large, requiring careful tuning and precise data segmentation. In this work, we present a method that uses a time window as the initial segmentation, then separates the data based on the change in the sensor value. Our system uses a multi-phase segmentation method that pulls all peaks and valleys for each axis of an accelerometer placed on the swimmer’s lower back. This results in high recognition performance using leave-one-subject-out validation on our study with 20 beginner swimmers, with our model optimized from our final dataset resulting in an F-Score of 0.95.

Keywords: Time window, peak/valley segmentation, feature extraction, beginner swimming, activity recognition.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 205

Authors: Chongya Song, Kang Yen, Alexander Pons, Jin Liu

Abstract:

In the field of machine learning, the ensemble has been employed as a common methodology to improve the performance upon multiple base classifiers. However, the true predictions are often canceled out by the false ones during consensus due to a phenomenon called “curse of correlation” which is represented as the strong interferences among the predictions produced by the base classifiers. In addition, the existing practices are still not able to effectively mitigate the problem of imbalanced classification. Based on the analysis on our experiment results, we conclude that the two problems are caused by some inherent deficiencies in the approach of consensus. Therefore, we create an enhanced ensemble algorithm which adopts a designed rank-based chain-mode consensus to overcome the two problems. In order to evaluate the proposed ensemble algorithm, we employ a well-known benchmark data set NSL-KDD (the improved version of dataset KDDCup99 produced by University of New Brunswick) to make comparisons between the proposed and 8 common ensemble algorithms. Particularly, each compared ensemble classifier uses the same 22 base classifiers, so that the differences in terms of the improvements toward the accuracy and reliability upon the base classifiers can be truly revealed. As a result, the proposed rank-based chain-mode consensus is proved to be a more effective ensemble solution than the traditional consensus approach, which outperforms the 8 ensemble algorithms by 20% on almost all compared metrices which include accuracy, precision, recall, F1-score and area under receiver operating characteristic curve.

Keywords: Consensus, curse of correlation, imbalanced classification, rank-based chain-mode ensemble.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 734

Authors: Jack R. McKenzie, Peter A. Appleby, Thomas House, Neil Walton

Abstract:

Cold-start is a notoriously difficult problem which can occur in recommendation systems, and arises when there is insufficient information to draw inferences for users or items. To address this challenge, a contextual bandit algorithm – the Fast Approximate Bayesian Contextual Cold Start Learning algorithm (FAB-COST) – is proposed, which is designed to provide improved accuracy compared to the traditionally used Laplace approximation in the logistic contextual bandit, while controlling both algorithmic complexity and computational cost. To this end, FAB-COST uses a combination of two moment projection variational methods: Expectation Propagation (EP), which performs well at the cold start, but becomes slow as the amount of data increases; and Assumed Density Filtering (ADF), which has slower growth of computational cost with data size but requires more data to obtain an acceptable level of accuracy. By switching from EP to ADF when the dataset becomes large, it is able to exploit their complementary strengths. The empirical justification for FAB-COST is presented, and systematically compared to other approaches on simulated data. In a benchmark against the Laplace approximation on real data consisting of over 670, 000 impressions from autotrader.co.uk, FAB-COST demonstrates at one point increase of over 16% in user clicks. On the basis of these results, it is argued that FAB-COST is likely to be an attractive approach to cold-start recommendation systems in a variety of contexts.

Keywords: Cold-start, expectation propagation, multi-armed bandits, Thompson sampling, variational inference.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 552

Authors: Mohd Hasmadi Ismail, Kamaruzaman Jusoff

Abstract:

In order to develop forest management strategies in tropical forest in Malaysia, surveying the forest resources and monitoring the forest area affected by logging activities is essential. There are tremendous effort has been done in classification of land cover related to forest resource management in this country as it is a priority in all aspects of forest mapping using remote sensing and related technology such as GIS. In fact classification process is a compulsory step in any remote sensing research. Therefore, the main objective of this paper is to assess classification accuracy of classified forest map on Landsat TM data from difference number of reference data (200 and 388 reference data). This comparison was made through observation (200 reference data), and interpretation and observation approaches (388 reference data). Five land cover classes namely primary forest, logged over forest, water bodies, bare land and agricultural crop/mixed horticultural can be identified by the differences in spectral wavelength. Result showed that an overall accuracy from 200 reference data was 83.5 % (kappa value 0.7502459; kappa variance 0.002871), which was considered acceptable or good for optical data. However, when 200 reference data was increased to 388 in the confusion matrix, the accuracy slightly improved from 83.5% to 89.17%, with Kappa statistic increased from 0.7502459 to 0.8026135, respectively. The accuracy in this classification suggested that this strategy for the selection of training area, interpretation approaches and number of reference data used were importance to perform better classification result.

Keywords: Image Classification, Reference Data, Accuracy Assessment, Kappa Statistic, Forest Land Cover

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

Authors: Mais Nijim, Rama Devi Chennuboyina, Waseem Al Aqqad

Abstract:

Advances in spatial and spectral resolution of satellite images have led to tremendous growth in large image databases. The data we acquire through satellites, radars, and sensors consists of important geographical information that can be used for remote sensing applications such as region planning, disaster management. Spatial data classification and object recognition are important tasks for many applications. However, classifying objects and identifying them manually from images is a difficult task. Object recognition is often considered as a classification problem, this task can be performed using machine-learning techniques. Despite of many machine-learning algorithms, the classification is done using supervised classifiers such as Support Vector Machines (SVM) as the area of interest is known. We proposed a classification method, which considers neighboring pixels in a region for feature extraction and it evaluates classifications precisely according to neighboring classes for semantic interpretation of region of interest (ROI). A dataset has been created for training and testing purpose; we generated the attributes by considering pixel intensity values and mean values of reflectance. We demonstrated the benefits of using knowledge discovery and data-mining techniques, which can be on image data for accurate information extraction and classification from high spatial resolution remote sensing imagery.

Keywords: Remote sensing, object recognition, classification, data mining, waterbody identification, feature extraction.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2055

Authors: L. Yu, W. K. Li, S. K. Ong, A. Y. C. Nee

Abstract:

In this paper, a scalable augmented reality framework for handheld devices is presented. The presented framework is enabled by using a server-client data communication structure, in which the search for tracking targets among a database of images is performed on the server-side while pixel-wise 3D tracking is performed on the client-side, which, in this case, is a handheld mobile device. Image search on the server-side adopts a residual-enhanced image descriptors representation that gives the framework a scalability property. The tracking algorithm on the client-side is based on a gravity-aligned feature descriptor which takes the advantage of a sensor-equipped mobile device and an optimized intensity-based image alignment approach that ensures the accuracy of 3D tracking. Automatic content streaming is achieved by using a key-frame selection algorithm, client working phase monitoring and standardized rules for content communication between the server and client. The recognition accuracy test performed on a standard dataset shows that the method adopted in the presented framework outperforms the Bag-of-Words (BoW) method that has been used in some of the previous systems. Experimental test conducted on a set of video sequences indicated the real-time performance of the tracking system with a frame rate at 15-30 frames per second. The presented framework is exposed to be functional in practical situations with a demonstration application on a campus walk-around.

Keywords: Augmented reality framework, server-client model, vision-based tracking, image search.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1140

Authors: K. Venkateswari, U. K. Balaji Saravanan, K. Thangaraj, K. V. Deepana

Abstract:

Image search engines rely on the surrounding textual keywords for the retrieval of images. It is a tedious work for the search engines like Google and Bing to interpret the user’s search intention and to provide the desired results. The recent researches also state that the Google image search engines do not work well on all the images. Consequently, this leads to the emergence of efficient image retrieval technique, which interprets the user’s search intention and shows the desired results. In order to accomplish this task, an efficient image re-ranking framework is required. Sequentially, to provide best image retrieval, the new image re-ranking framework is experimented in this paper. The implemented new image re-ranking framework provides best image retrieval from the image dataset by making use of re-ranking of retrieved images that is based on the user’s desired images. This is experimented in two sections. One is offline section and other is online section. In offline section, the reranking framework studies differently (reference classes or Semantic Spaces) for diverse user query keywords. The semantic signatures get generated by combining the textual and visual features of the images. In the online section, images are re-ranked by comparing the semantic signatures that are obtained from the reference classes with the user specified image query keywords. This re-ranking methodology will increases the retrieval image efficiency and the result will be effective to the user.

Keywords: CBIR, Image Re-ranking, Image Retrieval, Semantic Signature, Semantic Space.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1938

Authors: S. Hutasavi, D. Chen

Abstract:

The built-up area is a significant proxy to measure regional economic growth and reflects the Gross Provincial Product (GPP). However, an up-to-date and reliable database of built-up areas is not always available, especially in developing countries. The cloud-based geospatial analysis platform such as Google Earth Engine (GEE) provides an opportunity with accessibility and computational power for those countries to generate the built-up data. Therefore, this study aims to extract the built-up areas in Eastern Economic Corridor (EEC), Thailand using day and nighttime satellite imagery based on GEE facilities. The normalized indices were generated from Landsat 8 surface reflectance dataset, including Normalized Difference Built-up Index (NDBI), Built-up Index (BUI), and Modified Built-up Index (MBUI). These indices were applied to identify built-up areas in EEC. The result shows that MBUI performs better than BUI and NDBI, with the highest accuracy of 0.85 and Kappa of 0.82. Moreover, the overall accuracy of classification was improved from 79% to 90%, and error of total built-up area was decreased from 29% to 0.7%, after night-time light data from the Visible and Infrared Imaging Suite (VIIRS) Day Night Band (DNB). The results suggest that MBUI with night-time light imagery is appropriate for built-up area extraction and be utilize for further study of socioeconomic impacts of regional development policy over the EEC region.

Keywords: Built-up area extraction, Google earth engine, adaptive thresholding method, rapid mapping.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 610

Authors: Zixian Zhang, Shanliang Yao, Zile Huang, Zhaodong Wu, Xiaohui Zhu, Yong Yue, Jieming Ma

Abstract:

Unmanned Surface Vehicles (USVs) hold significant value for their capacity to undertake hazardous and labor-intensive operations over aquatic environments. Object detection tasks are significant in these applications. Nonetheless, the efficacy of USVs in object detection is impeded by several intrinsic challenges, including the intricate dispersal of obstacles, reflections emanating from coastal structures, and the presence of fog over water surfaces, among others. To address these problems, this paper provides a fusion method for USVs to effectively detect objects in the inland surface environment, utilizing vision sensors and 3+1D Millimeter-wave radar. The MMW radar is a complementary tool to vision sensors, offering reliable environmental data. This approach involves the conversion of the radar’s 3D point cloud into a 2D radar pseudo-image, thereby standardizing the format for radar and vision data by leveraging a point transformer. Furthermore, this paper proposes the development of a multi-source object detection network, named RV-YOLOX, which leverages radar-vision integration specifically tailored for inland waterway environments. The performance is evaluated on our self-recording waterways dataset. Compared with the YOLOX network, our fusion network significantly improves detection accuracy, especially for objects with bad light conditions.

Keywords: Inland waterways, object detection, YOLO, sensor fusion, self-attention, deep learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 295

Authors: Deokho Kim, Wonwoo Lee, Jaewoong Lee, Teresa Ng, Gun-Ill Lee, Jiwon Jeong

Abstract:

This paper presents a method for improving object search accuracy using a deep learning model. A major limitation to provide accurate similarity with deep learning is the requirement of huge amount of data for training pairwise similarity scores (metrics), which is impractical to collect. Thus, similarity scores are usually trained with a relatively small dataset, which comes from a different domain, causing limited accuracy on measuring similarity. For this reason, this paper proposes a deep learning model that can be trained with a significantly small amount of data, a clustered data which of each cluster contains a set of visually similar images. In order to measure similarity distance with the proposed method, visual features of two images are extracted from intermediate layers of a convolutional neural network with various pooling methods, and the network is trained with pairwise similarity scores which is defined zero for images in identical cluster. The proposed method outperforms the state-of-the-art object similarity scoring techniques on evaluation for finding exact items. The proposed method achieves 86.5% of accuracy compared to the accuracy of the state-of-the-art technique, which is 59.9%. That is, an exact item can be found among four retrieved images with an accuracy of 86.5%, and the rest can possibly be similar products more than the accuracy. Therefore, the proposed method can greatly reduce the amount of training data with an order of magnitude as well as providing a reliable similarity metric.

Keywords: Visual search, deep learning, convolutional neural network, machine learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 827

Authors: Jiakai Li, Gursel Serpen, Steven Selman, Matt Franchetti, Mike Riesen, Cynthia Schneider

Abstract:

This paper presents the development of a Bayesian belief network classifier for prediction of graft status and survival period in renal transplantation using the patient profile information prior to the transplantation. The objective was to explore feasibility of developing a decision making tool for identifying the most suitable recipient among the candidate pool members. The dataset was compiled from the University of Toledo Medical Center Hospital patients as reported to the United Network Organ Sharing, and had 1228 patient records for the period covering 1987 through 2009. The Bayes net classifiers were developed using the Weka machine learning software workbench. Two separate classifiers were induced from the data set, one to predict the status of the graft as either failed or living, and a second classifier to predict the graft survival period. The classifier for graft status prediction performed very well with a prediction accuracy of 97.8% and true positive values of 0.967 and 0.988 for the living and failed classes, respectively. The second classifier to predict the graft survival period yielded a prediction accuracy of 68.2% and a true positive rate of 0.85 for the class representing those instances with kidneys failing during the first year following transplantation. Simulation results indicated that it is feasible to develop a successful Bayesian belief network classifier for prediction of graft status, but not the graft survival period, using the information in UNOS database.

Keywords: Bayesian network classifier, renal transplantation, graft survival period, United Network for Organ Sharing

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2109

Authors: SP. Chokkalingam, K. Komathy

Abstract:

Advances in the field of image processing envision a new era of evaluation techniques and application of procedures in various different fields. One such field being considered is the biomedical field for prognosis as well as diagnosis of diseases. This plethora of methods though provides a wide range of options to select from, it also proves confusion in selecting the apt process and also in finding which one is more suitable. Our objective is to use a series of techniques on bone scans, so as to detect the occurrence of rheumatoid arthritis (RA) as accurately as possible. Amongst other techniques existing in the field our proposed system tends to be more effective as it depends on new methodologies that have been proved to be better and more consistent than others. Computer aided diagnosis will provide more accurate and infallible rate of consistency that will help to improve the efficiency of the system. The image first undergoes histogram smoothing and specification, morphing operation, boundary detection by edge following algorithm and finally image subtraction to determine the presence of rheumatoid arthritis in a more efficient and effective way. Using preprocessing noises are removed from images and using segmentation, region of interest is found and Histogram smoothing is applied for a specific portion of the images. Gray level co-occurrence matrix (GLCM) features like Mean, Median, Energy, Correlation, Bone Mineral Density (BMD) and etc. After finding all the features it stores in the database. This dataset is trained with inflamed and noninflamed values and with the help of neural network all the new images are checked properly for their status and Rough set is implemented for further reduction.

Keywords: Computer Aided Diagnosis, Edge Detection, Histogram Smoothing, Rheumatoid Arthritis.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2479

Authors: S. M. Seyedhosseini, A. Makui, M. Ghadami

Abstract:

Globalization and therefore increasing tight competition among companies, have resulted to increase the importance of making well-timed decision. Devising and employing effective strategies, that are flexible and adaptive to changing market, stand a greater chance of being effective in the long-term. In other side, a clear focus on managing the entire product lifecycle has emerged as critical areas for investment. Therefore, applying wellorganized tools to employ past experience in new case, helps to make proper and managerial decisions. Case based reasoning (CBR) is based on a means of solving a new problem by using or adapting solutions to old problems. In this paper, an adapted CBR model with k-nearest neighbor (K-NN) is employed to provide suggestions for better decision making which are adopted for a given product in the middle of life phase. The set of solutions are weighted by CBR in the principle of group decision making. Wrapper approach of genetic algorithm is employed to generate optimal feature subsets. The dataset of the department store, including various products which are collected among two years, have been used. K-fold approach is used to evaluate the classification accuracy rate. Empirical results are compared with classical case based reasoning algorithm which has no special process for feature selection, CBR-PCA algorithm based on filter approach feature selection, and Artificial Neural Network. The results indicate that the predictive performance of the model, compare with two CBR algorithms, in specific case is more effective.

Keywords: Case based reasoning, Genetic algorithm, Groupdecision making, Product management.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2174

Authors: Ajab G. Majidi, Bilal Bingölbali, Adem Akpınar

Abstract:

This study aims to investigate the amount of energy (economic wave energy potential) that can be obtained from the existing wave energy converters in the high wave energy potential region of the Black Sea in terms of wave energy potential and their performance at different depths in the region. The data needed for this purpose were obtained using the calibrated nested layered SWAN wave modeling program version 41.01AB, which was forced with Climate Forecast System Reanalysis (CFSR) winds from 1979 to 2009. The wave dataset at a time interval of 2 hours was accumulated for a sub-grid domain for around Karaburun beach in Arnavutkoy, a district of Istanbul city. The annual sea state characteristic matrices for the five different depths along with a vertical line to the coastline were calculated for 31 years. According to the power matrices of different wave energy converter systems and characteristic matrices for each possible installation depth, the probability distribution tables of the specified mean wave period or wave energy period and significant wave height were calculated. Then, by using the relationship between these distribution tables, according to the present wave climate, the energy that the wave energy converter systems at each depth can produce was determined. Thus, the economically feasible potential of the relevant coastal zone was revealed, and the effect of different depths on energy converter systems is presented. The Oceantic at 50, 75 and 100 m depths and Oyster at 5 and 25 m depths presents the best performance. In the 31-year long period 1998 the most and 1989 is the least dynamic year.

Keywords: Annual power production, Black Sea, efficiency, power production performance, wave energy converter.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 660

Authors: Mirjana Ruppel, Rajendra Persad, Amit Bahl, Sanja Dogramadzi, Chris Melhuish, Lyndon Smith

Abstract:

Multimodal image registration is a profoundly complex task which is why deep learning has been used widely to address it in recent years. However, two main challenges remain: Firstly, the lack of ground truth data calls for an unsupervised learning approach, which leads to the second challenge of defining a feasible loss function that can compare two images of different modalities to judge their level of alignment. To avoid this issue altogether we implement a generative adversarial network consisting of two registration networks GAB, GBA and two discrimination networks DA, DB connected by spatial transformation layers. GAB learns to generate a deformation field which registers an image of the modality B to an image of the modality A. To do that, it uses the feedback of the discriminator DB which is learning to judge the quality of alignment of the registered image B. GBA and DA learn a mapping from modality A to modality B. Additionally, a cycle-consistency loss is implemented. For this, both registration networks are employed twice, therefore resulting in images ˆA, ˆB which were registered to ˜B, ˜A which were registered to the initial image pair A, B. Thus the resulting and initial images of the same modality can be easily compared. A dataset of liver CT and MRI was used to evaluate the quality of our approach and to compare it against learning and non-learning based registration algorithms. Our approach leads to dice scores of up to 0.80 ± 0.01 and is therefore comparable to and slightly more successful than algorithms like SimpleElastix and VoxelMorph.

Keywords: Multimodal image registration, GAN, cycle consistency, deep learning.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 810

Authors: S. Boria

Abstract:

In the last years, the crashworthiness of an automotive body structure can be improved, since the beginning of the design stage, thanks to the development of specific optimization tools. It is well known how the finite element codes can help the designer to investigate the crashing performance of structures under dynamic impact. Therefore, by coupling nonlinear mathematical programming procedure and statistical techniques with FE simulations, it is possible to optimize the design with reduced number of analytical evaluations. In engineering applications, many optimization methods which are based on statistical techniques and utilize estimated models, called meta-models, are quickly spreading. A meta-model is an approximation of a detailed simulation model based on a dataset of input, identified by the design of experiments (DOE); the number of simulations needed to build it depends on the number of variables. Among the various types of meta-modeling techniques, Kriging method seems to be excellent in accuracy, robustness and efficiency compared to other ones when applied to crashworthiness optimization. Therefore the application of such meta-model was used in this work, in order to improve the structural optimization of a bumper for a racing car in composite material subjected to frontal impact. The specific energy absorption represents the objective function to maximize and the geometrical parameters subjected to some design constraints are the design variables. LS-DYNA codes were interfaced with LS-OPT tool in order to find the optimized solution, through the use of a domain reduction strategy. With the use of the Kriging meta-model the crashworthiness characteristic of the composite bumper was improved.

Keywords: Composite material, crashworthiness, finite element analysis, optimization.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1130

Authors: Abobaker Mohammed Qasem Farhan, Shangming Yang, Mohammed Al-Nehari

Abstract:

Pneumonia is a kind of lung disease that creates congestion in the chest. Such pneumonic conditions lead to loss of life due to the severity of high congestion. Pneumonic lung disease is caused by viral pneumonia, bacterial pneumonia, or COVID-19 induced pneumonia. The early prediction and classification of such lung diseases help reduce the mortality rate. We propose the automatic Computer-Aided Diagnosis (CAD) system in this paper using the deep learning approach. The proposed CAD system takes input from raw computerized tomography (CT) scans of the patient's chest and automatically predicts disease classification. We designed the Hybrid Deep Learning Algorithm (HDLA) to improve accuracy and reduce processing requirements. The raw CT scans are pre-processed first to enhance their quality for further analysis. We then applied a hybrid model that consists of automatic feature extraction and classification. We propose the robust 2D Convolutional Neural Network (CNN) model to extract the automatic features from the pre-processed CT image. This CNN model assures feature learning with extremely effective 1D feature extraction for each input CT image. The outcome of the 2D CNN model is then normalized using the Min-Max technique. The second step of the proposed hybrid model is related to training and classification using different classifiers. The simulation outcomes using the publicly available dataset prove the robustness and efficiency of the proposed model compared to state-of-art algorithms.

Keywords: CT scans, COVID-19, deep learning, image processing, pneumonia, lung disease.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 610

Authors: Peifei Zhu, Zisheng Li, Yasuki Kakishita, Mayumi Suzuki, Tomoaki Chono

Abstract:

Segmentation of left ventricle (LV) from cardiac ultrasound images provides a quantitative functional analysis of the heart to diagnose disease. Active Shape Model (ASM) is widely used for LV segmentation, but it suffers from the drawback that initialization of the shape model is not sufficiently close to the target, especially when dealing with abnormal shapes in disease. In this work, a two-step framework is improved to achieve a fast and efficient LV segmentation. First, a robust and efficient detection based on Hough forest localizes cardiac feature points. Such feature points are used to predict the initial fitting of the LV shape model. Second, ASM is applied to further fit the LV shape model to the cardiac ultrasound image. With the robust initialization, ASM is able to achieve more accurate segmentation. The performance of the proposed method is evaluated on a dataset of 810 cardiac ultrasound images that are mostly abnormal shapes. This proposed method is compared with several combinations of ASM and existing initialization methods. Our experiment results demonstrate that accuracy of the proposed method for feature point detection for initialization was 40% higher than the existing methods. Moreover, the proposed method significantly reduces the number of necessary ASM fitting loops and thus speeds up the whole segmentation process. Therefore, the proposed method is able to achieve more accurate and efficient segmentation results and is applicable to unusual shapes of heart with cardiac diseases, such as left atrial enlargement.

Keywords: Hough forest, active shape model, segmentation, cardiac left ventricle.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1504

Authors: Jameela Ali Alkrimi, Hamid A. Jalab, Loay E. George, Abdul Rahim Ahmad, Azizah Suliman, Karim Al-Jashamy

Abstract:

Red blood cells (RBC) are the most common types of blood cells and are the most intensively studied in cell biology. The lack of RBCs is a condition in which the amount of hemoglobin level is lower than normal and is referred to as “anemia”. Abnormalities in RBCs will affect the exchange of oxygen. This paper presents a comparative study for various techniques for classifying the RBCs as normal or abnormal (anemic) using WEKA. WEKA is an open source consists of different machine learning algorithms for data mining applications. The algorithms tested are Radial Basis Function neural network, Support vector machine, and K-Nearest Neighbors algorithm. Two sets of combined features were utilized for classification of blood cells images. The first set, exclusively consist of geometrical features, was used to identify whether the tested blood cell has a spherical shape or non-spherical cells. While the second set, consist mainly of textural features was used to recognize the types of the spherical cells. We have provided an evaluation based on applying these classification methods to our RBCs image dataset which were obtained from Serdang Hospital - Malaysia, and measuring the accuracy of test results. The best achieved classification rates are 97%, 98%, and 79% for Support vector machines, Radial Basis Function neural network, and K-Nearest Neighbors algorithm respectively.

Keywords: K-Nearest Neighbors, Neural Network, Radial Basis Function, Red blood cells, Support vector machine.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 2995

Authors: Mrinmoy Dhara, Vivek K. Sengar, Shovan L. Chattoraj, Soumiya Bhattacharjee

Abstract:

Remote sensing techniques have emerged as an asset for various geological studies. Satellite images obtained by different sensors contain plenty of information related to the terrain. Digital image processing further helps in customized ways for the prospecting of minerals. In this study, an attempt has been made to map the hydrothermally altered zones using multispectral and hyperspectral datasets of South East Rajasthan. Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER) and Hyperion (Level1R) dataset have been processed to generate different Band Ratio Composites (BRCs). For this study, ASTER derived BRCs were generated to delineate the alteration zones, gossans, abundant clays and host rocks. ASTER and Hyperion images were further processed to extract mineral end members and classified mineral maps have been produced using Spectral Angle Mapper (SAM) method. Results were validated with the geological map of the area which shows positive agreement with the image processing outputs. Thus, this study concludes that the band ratios and image processing in combination play significant role in demarcation of alteration zones which may provide pathfinders for mineral prospecting studies.

Keywords: Advanced space-borne thermal emission and reflection radiometer, ASTER, Hyperion, Band ratios, Alteration zones, spectral angle mapper.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1486

Authors: Akinola Ikudayisi, Josiah Adeyemo

Abstract:

The correct estimation of reference evapotranspiration (ETₒ) is required for effective irrigation water resources planning and management. However, there are some variables that must be considered while estimating and modeling ETₒ. This study therefore determines the multivariate analysis of correlated variables involved in the estimation and modeling of ETₒ at Vaalharts irrigation scheme (VIS) in South Africa using Principal Component Analysis (PCA) technique. Weather and meteorological data between 1994 and 2014 were obtained both from South African Weather Service (SAWS) and Agricultural Research Council (ARC) in South Africa for this study. Average monthly data of minimum and maximum temperature (°C), rainfall (mm), relative humidity (%), and wind speed (m/s) were the inputs to the PCA-based model, while ETₒ is the output. PCA technique was adopted to extract the most important information from the dataset and also to analyze the relationship between the five variables and ETₒ. This is to determine the most significant variables affecting ETₒ estimation at VIS. From the model performances, two principal components with a variance of 82.7% were retained after the eigenvector extraction. The results of the two principal components were compared and the model output shows that minimum temperature, maximum temperature and windspeed are the most important variables in ETₒ estimation and modeling at VIS. In order words, ETₒ increases with temperature and windspeed. Other variables such as rainfall and relative humidity are less important and cannot be used to provide enough information about ETₒ estimation at VIS. The outcome of this study has helped to reduce input variable dimensionality from five to the three most significant variables in ETₒ modelling at VIS, South Africa.

Keywords: Irrigation, principal component analysis, reference evapotranspiration, Vaalharts.

Procedia APA BibTeX Chicago EndNote Harvard JSON MLA RIS XML ISO 690 PDF Downloads 1061