Search results for: classification algorithms

Authors: Aydin Teymourifar, Gurkan Ozturk

Abstract:

In this paper, a hybrid distributed algorithm has been suggested for multi-objective dynamic flexible job shop scheduling problem. The proposed algorithm is high level, in which several algorithms search the space on different machines simultaneously also it is a hybrid algorithm that takes advantages of the artificial intelligence, evolutionary and optimization methods. Distribution is done at different levels and new approaches are used for design of the algorithm. Apache spark and Hadoop frameworks have been used for the distribution of the algorithm. The Pareto optimality approach is used for solving the multi-objective benchmarks. The suggested algorithm that is able to solve large-size problems in short times has been compared with the successful algorithms of the literature. The results prove high speed and efficiency of the algorithm.

Keywords: distributed algorithms, apache-spark, Hadoop, flexible dynamic job shop scheduling, multi-objective optimization

Procedia PDF Downloads 323

Authors: Steffy Maria Joseph, Abdu Rahiman V, Abdul Hameed K. M.

Abstract:

Online handwritten character recognition is a challenging field in Artificial Intelligence. The classification success rate of current techniques decreases when the dataset involves similarity and complexity in stroke styles, number of strokes and stroke characteristics variations. Malayalam is a complex south indian language spoken by about 35 million people especially in Kerala and Lakshadweep islands. In this paper, we consider the significant feature extraction for the similar stroke styles of Malayalam. This extracted feature set are suitable for the recognition of other handwritten south indian languages like Tamil, Telugu and Kannada. A classification scheme based on support vector machines (SVM) is proposed to improve the accuracy in classification and recognition of online malayalam handwritten characters. SVM Classifiers are the best for real world applications. The contribution of various features towards the accuracy in recognition is analysed. Performance for different kernels of SVM are also studied. A graphical user interface has developed for reading and displaying the character. Different writing styles are taken for each of the 44 alphabets. Various features are extracted and used for classification after the preprocessing of input data samples. Highest recognition accuracy of 97% is obtained experimentally at the best feature combination with polynomial kernel in SVM.

Keywords: SVM, matlab, malayalam, South Indian scripts, onlinehandwritten character recognition

Procedia PDF Downloads 552

Authors: Shelby Simpson, William Stanley, Namir Naba, Xiaodi Wang

Abstract:

Both t-SNE and UMAP are brand new state of art tools to predominantly preserve the local structure that is to group neighboring data points together, which indeed provides a very informative visualization of heterogeneity in our data. In this research, we develop a t-SNE and UMAP base neural network image classification algorithm to embed the original dataset to a corresponding low dimensional dataset as a preprocessing step, then use this embedded database as input to our specially designed neural network classifier for image classification. We use the fashion MNIST data set, which is a labeled data set of images of clothing objects in our experiments. t-SNE and UMAP are used for dimensionality reduction of the data set and thus produce low dimensional embeddings. Furthermore, we use the embeddings from t-SNE and UMAP to feed into two neural networks. The accuracy of the models from the two neural networks is then compared to a dense neural network that does not use embedding as an input to show which model can classify the images of clothing objects more accurately.

Keywords: t-SNE, UMAP, fashion MNIST, neural networks

Procedia PDF Downloads 168

Authors: Bahareh Golchin, Nooshin Riahi

Abstract:

One of the most significant issues as attended a lot in recent years is that of recognizing the sentiments and emotions in social media texts. The analysis of sentiments and emotions is intended to recognize the conceptual information such as the opinions, feelings, attitudes and emotions of people towards the products, services, organizations, people, topics, events and features in the written text. These indicate the greatness of the problem space. In the real world, businesses and organizations are always looking for tools to gather ideas, emotions, and directions of people about their products, services, or events related to their own. This article uses the Twitter social network, one of the most popular social networks with about 420 million active users, to extract data. Using this social network, users can share their information and opinions about personal issues, policies, products, events, etc. It can be used with appropriate classification of emotional states due to the availability of its data. In this study, supervised learning and deep neural network algorithms are used to classify the emotional states of Twitter users. The use of deep learning methods to increase the learning capacity of the model is an advantage due to the large amount of available data. Tweets collected on various topics are classified into four classes using a combination of two Bidirectional Long Short Term Memory network and a Convolutional network. The results obtained from this study with an average accuracy of 93%, show good results extracted from the proposed framework and improved accuracy compared to previous work.

Keywords: emotion classification, sentiment analysis, social networks, deep neural networks

Procedia PDF Downloads 114

Authors: Ali J. Ghandour, Houssam A. Krayem, Abedelkarim A. Jezzini

Abstract:

High-resolution satellite images and remote sensing can provide global information in a fast way compared to traditional methods of data collection. Under such high resolution, a road is not a thin line anymore. Objects such as cars and trees are easily identifiable. Automatic vehicles enumeration can be considered one of the most important applications in traffic management. In this paper, autonomous vehicle detection and classification approach in highway environment is proposed. This approach consists mainly of three stages: (i) first, a set of preprocessing operations are applied including soil, vegetation, water suppression. (ii) Then, road networks detection and delineation is implemented using built-up area index, followed by several morphological operations. This step plays an important role in increasing the overall detection accuracy since vehicles candidates are objects contained within the road networks only. (iii) Multi-level Otsu segmentation is implemented in the last stage, resulting in vehicle detection and classification, where detected vehicles are classified into cars and trucks. Accuracy assessment analysis is conducted over different study areas to show the great efficiency of the proposed method, especially in highway environment.

Keywords: remote sensing, object identification, vehicle and road extraction, vehicle and road features-based classification

Procedia PDF Downloads 207

Authors: Majid Habib Khan, Jinwei Zhao, Xinhong Hei, Liu Jiedong, Rana Shahzad Noor, Muhammad Imran

Abstract:

Deep learning is increasingly employed in image classification, yet the scarcity and high cost of labeled data for training remain a challenge. Limited samples often lead to overfitting due to biased sample distribution. This paper introduces a dynamic distribution calibration method for few-shot learning. Initially, base and new class samples undergo normalization to mitigate disparate feature magnitudes. A pre-trained model then extracts feature vectors from both classes. The method dynamically selects distribution characteristics from base classes (both adjacent and remote) in the embedding space, using a threshold value approach for new class samples. Given the propensity of similar classes to share feature distributions like mean and variance, this research assumes a Gaussian distribution for feature vectors. Subsequently, distributional features of new class samples are calibrated using a corrected hyperparameter, derived from the distribution features of both adjacent and distant base classes. This calibration augments the new class sample set. The technique demonstrates significant improvements, with up to 4% accuracy gains in few-shot classification challenges, as evidenced by tests on miniImagenet and CUB datasets.

Keywords: deep learning, computer vision, image classification, few-shot learning, threshold

Procedia PDF Downloads 37

Authors: Mekamı Hayet, Bounoua Nacer, Benabderrahmane Sidahmed, Taleb Ahmed

Abstract:

Pose estimation is an important task in computer vision. Though the majority of the existing solutions provide good accuracy results, they are often overly complex and computationally expensive. In this perspective, we propose the use of dimensionality reduction techniques to address the problem of facial pose estimation. Firstly, a face image is converted into one-dimensional time series using Hilbert space filling curve, then the approach converts these time series data to a symbolic representation. Furthermore, a distance matrix is calculated between symbolic series of an input learning dataset of images, to generate classifiers of frontal vs. profile face pose. The proposed method is evaluated with three public datasets. Experimental results have shown that our approach is able to achieve a correct classification rate exceeding 97% with K-NN algorithm.

Keywords: machine learning, pattern recognition, facial pose classification, time series

Procedia PDF Downloads 325

Authors: Kenan Morani, Esra Kaya Ayana

Abstract:

This study aimed to develop a novel pipeline for COVID-19 detection using a large and rigorously annotated database of computed tomography (CT) images. The pipeline consists of UNet-based segmentation, lung extraction, and a classification part, with the addition of optional slice removal techniques following the segmentation part. In this work, a batch normalization was added to the original UNet model to produce lighter and better localization, which is then utilized to build a full pipeline for COVID-19 diagnosis. To evaluate the effectiveness of the proposed pipeline, various segmentation methods were compared in terms of their performance and complexity. The proposed segmentation method with batch normalization outperformed traditional methods and other alternatives, resulting in a higher dice score on a publicly available dataset. Moreover, at the slice level, the proposed pipeline demonstrated high validation accuracy, indicating the efficiency of predicting 2D slices. At the patient level, the full approach exhibited higher validation accuracy and macro F1 score compared to other alternatives, surpassing the baseline. The classification component of the proposed pipeline utilizes a convolutional neural network (CNN) to make final diagnosis decisions. The COV19-CT-DB dataset, which contains a large number of CT scans with various types of slices and rigorously annotated for COVID-19 detection, was utilized for classification. The proposed pipeline outperformed many other alternatives on the dataset.

Keywords: classification, computed tomography, lung extraction, macro F1 score, UNet segmentation

Procedia PDF Downloads 106

Authors: Guoliang Lu, Changhou Lu, Xueyong Li

Abstract:

In action recognition, previous studies have demonstrated the effectiveness of using multiple features to improve the recognition performance. We focus on two practical issues: i) most studies use a direct way of concatenating/accumulating multi features to evaluate the similarity between two actions. This way could be too strong since each kind of feature can include different dimensions, quantities, etc; ii) in many studies, the employed classification methods lack of a flexible and effective mechanism to add new feature(s) into classification. In this paper, we explore an unified scheme based on recently-proposed multi-dimensional dynamic time warping (MD-DTW). Experiments demonstrated the scheme's effectiveness of combining multi-feature and the flexibility of adding new feature(s) to increase the recognition performance. In addition, the explored scheme also provides us an open architecture for using new advanced classification methods in the future to enhance action recognition.

Keywords: action recognition, multi features, dynamic time warping, feature combination

Procedia PDF Downloads 418

Authors: Christos Chaschatzis, Chrysoula Karaiskou, Pantelis Angelidis, Sotirios K. Goudos, Igor Kotsiuba, Panagiotis Sarigiannidis

Abstract:

Over the past few decades, the rapid growth of the global population has led to the need to increase agricultural production and improve the quality of agricultural goods. There is a growing focus on environmentally eco-friendly solutions, sustainable production, and biologically minimally fertilized products in contemporary society. Precision agriculture has the potential to incorporate a wide range of innovative solutions with the development of machine learning algorithms. YOLOv5 and YOLOv8 are two of the most advanced object detection algorithms capable of accurately recognizing objects in real time. Detecting tree diseases is crucial for improving the food production rate and ensuring sustainability. This research aims to evaluate the efficacy of YOLOv5 and YOLOv8 in detecting the symptoms of Armillaria spp. in sweet cherry trees and determining their health status, with the goal of enhancing the robustness of precision agriculture. Additionally, this study will explore Computer Vision (CV) techniques with machine learning algorithms to improve the detection process’s efficiency.

Keywords: Armillaria spp., machine learning, precision agriculture, smart farming, sweet cherries trees, YOLOv5, YOLOv8

Procedia PDF Downloads 82

Authors: Mohammad Shams Esfand Abadi, Mohammad Ranjbar, Reza Ebrahimpour

Abstract:

This work presents the distributed processing solution problem in a diffusion network based on the adapt then combine (ATC) and combine then adapt (CTA)selective partial update normalized least mean squares (SPU-NLMS) algorithms. Also, we extend this approach to dynamic selection affine projection algorithm (DS-APA) and ATC-DS-APA and CTA-DS-APA are established. The purpose of ATC-SPU-NLMS and CTA-SPU-NLMS algorithm is to reduce the computational complexity by updating the selected blocks of weight coefficients at every iteration. In CTA-DS-APA and ATC-DS-APA, the number of the input vectors is selected dynamically. Diffusion cooperation strategies have been shown to provide good performance based on these algorithms. The good performance of introduced algorithm is illustrated with various experimental results.

Keywords: selective partial update, affine projection, dynamic selection, diffusion, adaptive distributed networks

Procedia PDF Downloads 676

Authors: Vahid Bairami Rad

Abstract:

Extracting knowledge from spatial data like GIS data is important to reduce the data and extract information. Therefore, the development of new techniques and tools that support the human in transforming data into useful knowledge has been the focus of the relatively new and interdisciplinary research area ‘knowledge discovery in databases’. Thus, we introduce a set of database primitives or basic operations for spatial data mining which are sufficient to express most of the spatial data mining algorithms from the literature. This approach has several advantages. Similar to the relational standard language SQL, the use of standard primitives will speed-up the development of new data mining algorithms and will also make them more portable. We introduced a database-oriented framework for spatial data mining which is based on the concepts of neighborhood graphs and paths. A small set of basic operations on these graphs and paths were defined as database primitives for spatial data mining. Furthermore, techniques to efficiently support the database primitives by a commercial DBMS were presented.

Keywords: spatial data base, knowledge discovery database, data mining, spatial relationship, predictive data mining

Procedia PDF Downloads 430

Authors: Anukriti Kumar, Tanmay Singh, Dinesh Kumar Vishwakarma

Abstract:

Traffic control has been one of the most common and irritating problems since the time automobiles have hit the roads. Problems like traffic congestion have led to a significant time burden around the world and one significant solution to these problems can be the proper implementation of the Intelligent Transport System (ITS). It involves the integration of various tools like smart sensors, artificial intelligence, position technologies and mobile data services to manage traffic flow, reduce congestion and enhance driver's ability to avoid accidents during adverse weather. Road and traffic signs’ recognition is an emerging field of research in ITS. Classification problem of traffic signs needs to be solved as it is a major step in our journey towards building semi-autonomous/autonomous driving systems. The purpose of this work focuses on implementing an approach to solve the problem of traffic sign classification by developing a Convolutional Neural Network (CNN) classifier using the GTSRB (German Traffic Sign Recognition Benchmark) dataset. Rather than using hand-crafted features, our model addresses the concern of exploding huge parameters and data method augmentations. Our model achieved an accuracy of around 97.6% which is comparable to various state-of-the-art architectures.

Keywords: multiclass classification, convolution neural network, OpenCV

Procedia PDF Downloads 153

Authors: Ebtehal Aljedaani, Maha Aljohani

Abstract:

Privacy and security patterns are both important for developing software that protects users' data and privacy. Privacy patterns are designed to address common privacy problems, such as unauthorized data collection and disclosure. Security patterns are designed to protect software from attack and ensure reliability and trustworthiness. Using privacy and security patterns, software engineers can implement security and privacy by design principles, which means that security and privacy are considered throughout the software development process. These patterns are available to translate "security & privacy-by-design" into practical advice for software engineering. Previous research on privacy and security patterns has typically focused on one category of patterns at a time. This paper aims to bridge this gap by merging the two categories and identifying their similarities and differences. To do this, the authors conducted a systematic literature review of 25 research papers on privacy and security patterns. The papers were analysed based on the category of the pattern, the classification of the pattern, and the security requirements that the pattern addresses. This paper presents the results of a comprehensive review of privacy and security design patterns. The review is intended to help future IT designers understand the relationship between the two types of patterns and how to use them to design secure and privacy-preserving software. The paper provides a clear classification of privacy and security design patterns, along with examples of each type. The authors found that there is only one widely accepted classification of privacy design patterns, while there are several competing classifications of security design patterns. Three types of security design patterns were found to be the most commonly used.

Keywords: design patterns, security, privacy, classification of patterns, security patterns, privacy patterns

Procedia PDF Downloads 98

Authors: R. R. Ramsheeja, R. Sreeraj

Abstract:

For view the internal structures of the human body such as liver, brain, kidney etc have a wide range of different modalities for medical images are provided nowadays. Computer Tomography is one of the most significant medical image modalities. In this paper use CT liver images for study the use of automatic computer aided techniques to calculate the volume of the liver tumor. Segmentation method is used for the detection of tumor from the CT scan is proposed. Gaussian filter is used for denoising the liver image and Adaptive Thresholding algorithm is used for segmentation. Multiple Region Of Interest(ROI) based method that may help to characteristic the feature different. It provides a significant impact on classification performance. Due to the characteristic of liver tumor lesion, inherent difficulties appear selective. For a better performance, a novel proposed system is introduced. Multiple ROI based feature selection and classification are performed. In order to obtain of relevant features for Support Vector Machine(SVM) classifier is important for better generalization performance. The proposed system helps to improve the better classification performance, reason in which we can see a significant reduction of features is used. The diagnosis of liver cancer from the computer tomography images is very difficult in nature. Early detection of liver tumor is very helpful to save the human life.

Keywords: computed tomography (CT), multiple region of interest(ROI), feature values, segmentation, SVM classification

Procedia PDF Downloads 488

Authors: Doyin Afolabi, Phillip Adewole, Oladipupo Sennaike

Abstract:

Most well-known classifiers, including the decision tree algorithm, can make predictions on balanced datasets efficiently. However, the decision tree algorithm tends to be biased towards imbalanced datasets because of the skewness of the distribution of such datasets. To overcome this problem, this study proposes a weighted decision tree algorithm that aims to remove the bias toward the majority class and prevents the reduction of majority observations in imbalance datasets classification. The proposed weighted decision tree algorithm was tested on three imbalanced datasets- cancer dataset, german credit dataset, and banknote dataset. The specificity, sensitivity, and accuracy metrics were used to evaluate the performance of the proposed decision tree algorithm on the datasets. The evaluation results show that for some of the weights of our proposed decision tree, the specificity, sensitivity, and accuracy metrics gave better results compared to that of the ID3 decision tree and decision tree induced with minority entropy for all three datasets.

Keywords: data mining, decision tree, classification, imbalance dataset

Procedia PDF Downloads 100

Authors: Eiman Kattan

Abstract:

This study aims to evaluate the impact of classifying labelled remote sensing images conventional neural network (CNN) architecture, i.e., AlexNet on different land cover scenarios based on two remotely sensed datasets from different point of views such as the computational time and performance. Thus, a set of experiments were conducted to specify the effectiveness of the selected convolutional neural network using two implementing approaches, named fully trained and fine-tuned. For validation purposes, two remote sensing datasets, AID, and RSSCN7 which are publicly available and have different land covers features were used in the experiments. These datasets have a wide diversity of input data, number of classes, amount of labelled data, and texture patterns. A specifically designed interactive deep learning GPU training platform for image classification (Nvidia Digit) was employed in the experiments. It has shown efficiency in training, validation, and testing. As a result, the fully trained approach has achieved a trivial result for both of the two data sets, AID and RSSCN7 by 73.346% and 71.857% within 24 min, 1 sec and 8 min, 3 sec respectively. However, dramatic improvement of the classification performance using the fine-tuning approach has been recorded by 92.5% and 91% respectively within 24min, 44 secs and 8 min 41 sec respectively. The represented conclusion opens the opportunities for a better classification performance in various applications such as agriculture and crops remote sensing.

Keywords: conventional neural network, remote sensing, land cover, land use

Procedia PDF Downloads 343

Authors: Arian Hosseini, Mahmudul Hasan

Abstract:

To address the increasing need for efficient and accurate content moderation, we propose an efficient and lightweight deep classification ensemble structure. Our approach is based on a combination of simple visual features, designed for high-accuracy classification of violent content with low false positives. Our ensemble architecture utilizes a set of lightweight models with narrowed-down color features, and we apply it to both images and videos. We evaluated our approach using a large dataset of explosion and blast contents and compared its performance to popular deep learning models such as ResNet-50. Our evaluation results demonstrate significant improvements in prediction accuracy, while benefiting from 7.64x faster inference and lower computation cost. While our approach is tailored to explosion detection, it can be applied to other similar content moderation and violence detection use cases as well. Based on our experiments, we propose a "think small, think many" philosophy in classification scenarios. We argue that transforming a single, large, monolithic deep model into a verification-based step model ensemble of multiple small, simple, and lightweight models with narrowed-down visual features can possibly lead to predictions with higher accuracy.

Keywords: deep classification, content moderation, ensemble learning, explosion detection, video processing

Procedia PDF Downloads 24

Authors: Yulong Wang, Yuan Yan Tang, Cuiming Zou, Lina Yang

Abstract:

This paper presents a novel sparse representation method for robust pattern classification. Generalized orthogonal matching pursuit (GOMP) is a recently proposed efficient sparse representation technique. However, GOMP adopts the mean square error (MSE) criterion and assign the same weights to all measurements, including both severely and slightly corrupted ones. To reduce the limitation, we propose an information-theoretic GOMP (ITGOMP) method by exploiting the correntropy induced metric. The results show that ITGOMP can adaptively assign small weights on severely contaminated measurements and large weights on clean ones, respectively. An ITGOMP based classifier is further developed for robust pattern classification. The experiments on public real datasets demonstrate the efficacy of the proposed approach.

Keywords: correntropy induced metric, matching pursuit, pattern classification, sparse representation

Procedia PDF Downloads 333

Authors: Qi Xiu, Hiromu Hota, Yohsuke Ishii, Takuya Oda

Abstract:

Recently, collectable manufacturing data are rapidly increasing. On the other hand, mega recall is getting serious as a social problem. Under such circumstances, there are increasing needs for preventing mega recalls by defect analysis such as root cause analysis and abnormal detection utilizing manufacturing data. However, the time to classify strings in manufacturing data by traditional method is too long to meet requirement of quick defect analysis. Therefore, we present String Length Distribution Classification method (SLDC) to correctly classify strings in a short time. This method learns character features, especially string length distribution from Product ID, Machine ID in BOM and asset list. By applying the proposal to strings in actual manufacturing data, we verified that the classification time of strings can be reduced by 80%. As a result, it can be estimated that the requirement of quick defect analysis can be fulfilled.

Keywords: string classification, data quality, feature selection, probability distribution, string length

Procedia PDF Downloads 295

Authors: Samiah Alammari, Nassim Ammour

Abstract:

When providing a massive number of tasks successively to a deep learning process, a good performance of the model requires preserving the previous tasks data to retrain the model for each upcoming classification. Otherwise, the model performs poorly due to the catastrophic forgetting phenomenon. To overcome this shortcoming, we developed a successful continual learning deep model for remote sensing hyperspectral image regions classification. The proposed neural network architecture encapsulates two trainable subnetworks. The first module adapts its weights by minimizing the discrimination error between the land-cover classes during the new task learning, and the second module tries to learn how to replicate the data of the previous tasks by discovering the latent data structure of the new task dataset. We conduct experiments on HSI dataset Indian Pines. The results confirm the capability of the proposed method.

Keywords: continual learning, data reconstruction, remote sensing, hyperspectral image segmentation

Procedia PDF Downloads 220

Authors: Jularat Chumnaul

Abstract:

In forensic science, corpses from various homicides are different; there are both complete and incomplete, depending on causes of death or forms of homicide. For example, some corpses are cut into pieces, some are camouflaged by dumping into the river, some are buried, some are burned to destroy the evidence, and others. If the corpses are incomplete, it can lead to the difficulty of personally identifying because some tissues and bones are destroyed. To specify gender of the corpses from skeletal remains, the most precise method is DNA identification. However, this method is costly and takes longer so that other identification techniques are used instead. The first technique that is widely used is considering the features of bones. In general, an evidence from the corpses such as some pieces of bones, especially the skull and pelvis can be used to identify their gender. To use this technique, forensic scientists are required observation skills in order to classify the difference between male and female bones. Although this technique is uncomplicated, saving time and cost, and the forensic scientists can fairly accurately determine gender by using this technique (apparently an accuracy rate of 90% or more), the crucial disadvantage is there are only some positions of skeleton that can be used to specify gender such as supraorbital ridge, nuchal crest, temporal lobe, mandible, and chin. Therefore, the skeletal remains that will be used have to be complete. The other technique that is widely used for gender specifying in forensic science and archeology is skeletal measurements. The advantage of this method is it can be used in several positions in one piece of bones, and it can be used even if the bones are not complete. In this study, the classification and cluster analysis are applied to this technique, including the Kth Nearest Neighbor Classification, Classification Tree, Ward Linkage Cluster, K-mean Cluster, and Two Step Cluster. The data contains 507 particular individuals and 9 skeletal measurements (diameter measurements), and the performance of five methods are investigated by considering the apparent error rate (APER). The results from this study indicate that the Two Step Cluster and Kth Nearest Neighbor method seem to be suitable to specify gender from human skeletal remains because both yield small apparent error rate of 0.20% and 4.14%, respectively. On the other hand, the Classification Tree, Ward Linkage Cluster, and K-mean Cluster method are not appropriate since they yield large apparent error rate of 10.65%, 10.65%, and 16.37%, respectively. However, there are other ways to evaluate the performance of classification such as an estimate of the error rate using the holdout procedure or misclassification costs, and the difference methods can make the different conclusions.

Keywords: skeletal measurements, classification, cluster, apparent error rate

Procedia PDF Downloads 231

Authors: Evan Lowhorn, Rocio Alba-Flores

Abstract:

The purpose of this work is to develop a method for classifying hand signals and using the output in a drone control algorithm. To achieve this, methods based on Convolutional Neural Networks (CNN) were applied. CNN's are a subset of deep learning, which allows grid-like inputs to be processed and passed through a neural network to be trained for classification. This type of neural network allows for classification via imaging, which is less intrusive than previous methods using biosensors, such as EMG sensors. Classification CNN's operate purely from the pixel values in an image; therefore they can be used without additional exteroceptive sensors. A development bench was constructed using a desktop computer connected to a high-definition webcam mounted on a scissor arm. This allowed the camera to be pointed downwards at the desk to provide a constant solid background for the dataset and a clear detection area for the user. A MATLAB script was created to automate dataset image capture at the development bench and save the images to the desktop. This allowed the user to create their own dataset of 12,000 images within three hours. These images were evenly distributed among seven classes. The defined classes include forward, backward, left, right, idle, and land. The drone has a popular flip function which was also included as an additional class. To simplify control, the corresponding hand signals chosen were the numerical hand signs for one through five for movements, a fist for land, and the universal “ok” sign for the flip command. Transfer learning with PyTorch (Python) was performed using a pre-trained 18-layer residual learning network (ResNet-18) to retrain the network for custom classification. An algorithm was created to interpret the classification and send encoded messages to a Ryze Tello drone over its 2.4 GHz Wi-Fi connection. The drone’s movements were performed in half-meter distance increments at a constant speed. When combined with the drone control algorithm, the classification performed as desired with negligible latency when compared to the delay in the drone’s movement commands.

Keywords: classification, computer vision, convolutional neural networks, drone control

Procedia PDF Downloads 185

Authors: Farhat Imtiaz, Umar Farooq

Abstract:

In this work, we developed a computational model to predict ply level failure in impacted composite laminates. Data obtained from physical testing from flat and round nose impacts of 8-, 16-, 24-ply laminates were considered. Routine inspections of the tested laminates were carried out to approximate ply by ply inflicted damage incurred. Plots consisting of load–time, load–deflection, and energy–time history were drawn to approximate the inflicted damages. Impact test generated unwanted data logged due to restrictions on testing and logging systems were also filtered. Conventional filters (built-in, statistical, and numerical) reliably predicted load thresholds for relatively thin laminates such as eight and sixteen ply panels. However, for relatively thick laminates such as twenty-four ply laminates impacted by flat nose impact generated clipped data which can just be de-noised using oscillatory algorithms. The literature search reveals that modern oscillatory data filtering and extrapolation algorithms have scarcely been utilized. This investigation reports applications of filtering and extrapolation of the clipped data utilising fast Fourier Convolution algorithm to predict load thresholds. Some of the results were related to the impact-induced damage areas identified with Ultrasonic C-scans and found to be in acceptable agreement. Based on consistent findings, utilizing of modern data filtering and extrapolation algorithms to data logged by the existing machines has efficiently enhanced data interpretations without resorting to extra resources. The algorithms could be useful for impact-induced damage approximations of similar cases.

Keywords: fibre reinforced laminates, fast Fourier algorithms, mechanical testing, data filtering and extrapolation

Procedia PDF Downloads 115

Authors: Javier Alfonso Bonilla-Chávez, Angélica Lozano

Abstract:

In North America, more than 2,000 people annually die in accidents related to railroad tracks. In 2020, collisions at grade crossings were the main cause of deaths related to railway accidents in Mexico. Railway networks have constant interaction with motor transport users, cyclists, and pedestrians, mainly in grade crossings, where is the greatest vulnerability and risk of accidents. Usually, accidents at grade crossings are directly related to risky behavior and non-compliance with regulations by motorists, cyclists, and pedestrians, especially in developing countries. Around the world, countries classify these crossings in different ways. In Mexico, according to their dangerousness (high, medium, or low), types A, B and C have been established, recommending for each one different type of auditive and visual signaling and gates, as well as horizontal and vertical signaling. This classification is based in a weighting, but regrettably, it is not explained how the weight values were obtained. A review of the variables and the current approach for the grade crossing classification is required, since it is inadequate for some crossings. In contrast, North America (USA and Canada) and European countries consider a broader classification so that attention to each crossing is addressed more precisely and equipment costs are adjusted. Lack of a proper classification, could lead to cost overruns in the equipment and a deficient operation. To exemplify the lack of a good classification, six crossings are studied, three located in the rural area of Mexico and three in Mexico City. These cases show the need of: improving the current regulations, improving the existing infrastructure, and implementing technological systems, including informative signals with nomenclature of the involved crossing and direct telephone line for reporting emergencies. This implementation is unaffordable for most municipal governments. Also, an inventory of the most dangerous grade crossings in urban and rural areas must be obtained. Then, an approach for improving the classification of grade crossings is suggested. This approach must be based on criteria design, characteristics of adjacent roads or intersections which can influence traffic flow through the crossing, accidents related to motorized and non-motorized vehicles, land use and land management, type of area, and services and economic activities in the zone where the grade crossings is located. An expanded classification of grade crossing in Mexico could reduce accidents and improve the efficiency of the railroad.

Keywords: accidents, grade crossing, railroad, traffic safety

Procedia PDF Downloads 87

Authors: P. S. Jagadeesh Kumar, Yang Yung, Wenli Hu

Abstract:

Speech emotion classification is a dominant research field in finding a sturdy and profligate classifier appropriate for different real-life applications. This effort accentuates on classifying different emotions from speech signal quarried from the features related to pitch, formants, energy contours, jitter, shimmer, spectral, perceptual and temporal features. Tensor deep stacking neural networks were supported to examine the factors that influence the classification success rate. Facial electromyography signals were composed of several forms of focuses in a controlled atmosphere by means of audio-visual stimuli. Proficient facial electromyography signals were pre-processed using moving average filter, and a set of arithmetical features were excavated. Extracted features were mapped into consistent emotions using bilinear mapping. With facial electromyography signals, a database comprising diverse emotions will be exposed with a suitable fine-tuning of features and training data. A success rate of 92% can be attained deprived of increasing the system connivance and the computation time for sorting diverse emotional states.

Keywords: speech emotion classification, tensor deep stacking neural networks, facial electromyography, bilinear mapping, audio-visual stimuli

Procedia PDF Downloads 223

Authors: Li Kewen, Su Zhaoxin, Wang Xingmou, Zhu Jian Bing

Abstract:

Conventional reservoir prediction methods ar not sufficient to explore the implicit relation between seismic attributes, and thus data utilization is low. In order to improve the predictive classification accuracy of reservoir lithology, this paper proposes a deep learning lithology prediction method based on ResNet (Residual Neural Network) and SENet (Squeeze-and-Excitation Neural Network). The neural network model is built and trained by using seismic attribute data and lithology data of Shengli oilfield, and the nonlinear mapping relationship between seismic attribute and lithology marker is established. The experimental results show that this method can significantly improve the classification effect of reservoir lithology, and the classification accuracy is close to 70%. This study can effectively predict the lithology of undrilled area and provide support for exploration and development.

Keywords: convolutional neural network, lithology, prediction of reservoir, seismic attributes

Procedia PDF Downloads 150

Authors: Regina Chua

Abstract:

To reduce the costs of re-fielding a large survey, a Random Forest classifier was applied to measure the accuracy of classifying individuals into their assigned segments with the fewest possible questions. Given a long survey, one needed to determine the most predictive ten or fewer questions that would accurately assign new individuals to custom segments. Furthermore, the solution needed to be quick in its classification and usable in non-Python environments. In this paper, a supervised Random Forest classifier was modeled on a dataset with 7,000 individuals, 60 questions, and 254 features. The Random Forest consisted of an iterative collection of individual decision trees that result in a predicted segment with robust precision and recall scores compared to a single tree. A random 70-30 stratified sampling for training the algorithm was used, and accuracy trade-offs at different depths for each segment were identified. Ultimately, the Random Forest classifier performed at 87% accuracy at a depth of 10 with 20 instead of 254 features and 10 instead of 60 questions. With an acceptable accuracy in prioritizing feature selection, new tools were developed for non-Python environments: a worksheet with a formulaic version of the algorithm and an embedded function to predict the segment of an individual in real-time. Random Forest was determined to be an optimal classification model by its feature selection, performance, processing speed, and flexible application in other environments.

Keywords: machine learning, supervised learning, data science, random forest, classification, prediction, predictive modeling

Procedia PDF Downloads 73

Authors: Nicholas Milburn, Yu Liang, Dalei Wu

Abstract:

Athlete performance scoring within the climbing do-main presents interesting challenges as the sport does not have an objective way to assign skill. Assessing skill levels within any sport is valuable as it can be used to mark progress while training, and it can help an athlete choose appropriate climbs to attempt. Machine learning-based methods are popular for complex problems like this. The dataset available was composed of dynamic force data recorded during climbing; however, this dataset came with challenges such as data scarcity, imbalance, and it was temporally heterogeneous. Investigated solutions to these challenges include data augmentation, temporal normalization, conversion of time series to the spectral domain, and cross validation strategies. The investigated solutions to the classification problem included light weight machine classifiers KNN and SVM as well as the deep learning with CNN. The best performing model had an 80% accuracy. In conclusion, there seems to be enough information within climbing force data to accurately categorize climbers by skill.

Keywords: classification, climbing, data imbalance, data scarcity, machine learning, time sequence

Procedia PDF Downloads 122

Authors: M. Duran Toksari, Berrin Ucarkus

Abstract:

In this paper, we propose two algorithms to optimally solve makespan and total completion time scheduling problems with learning effect and job dependent delivery times in a single machine environment. The delivery time is the extra time to eliminate adverse effect between the main processing and delivery to the customer. In this paper, we introduce the job dependent delivery times for some single machine scheduling problems with position dependent learning effect, which are makespan are total completion. The results with respect to two algorithms proposed for solving of the each problem are compared with LINGO solutions for 50-jobs, 100-jobs and 150-jobs problems. The proposed algorithms can find the same results in shorter time.

Keywords: delivery Times, learning effect, makespan, scheduling, total completion time

Procedia PDF Downloads 452