Search results for: hyperspectral image classification using tree search algorithm

Authors: Khald Anisse, Amina Radgui, Mohammed Rziza

Abstract:

Tracking objects on video sequences is a very challenging task in many works in computer vision applications. However, there is no article that treats this topic in catadioptric vision. This paper is an attempt that tries to describe a new approach of omnidirectional images processing based on inverse stereographic projection in the half-sphere. We used the spherical model proposed by Gayer and al. For object tracking, our work is based on snake method, with optimization using the Greedy algorithm, by adapting its different operators. The algorithm will respect the deformed geometries of omnidirectional images such as spherical neighborhood, spherical gradient and reformulation of optimization algorithm on the spherical domain. This tracking method that we call "spherical snake" permitted to know the change of the shape and the size of object in different replacements in the spherical image.

Keywords: computer vision, spherical snake, omnidirectional image, object tracking, inverse stereographic projection

Procedia PDF Downloads 404

Authors: Sihem Oujdi, Hafida Belbachir

Abstract:

Existing methods of data mining cannot be applied on spatial data because they require spatial specificity consideration, as spatial relationships. This paper focuses on the classification with decision trees, which are one of the data mining techniques. We propose an extension of the C4.5 algorithm for spatial data, based on two different approaches Join materialization and Querying on the fly the different tables. Similar works have been done on these two main approaches, the first - Join materialization - favors the processing time in spite of memory space, whereas the second - Querying on the fly different tables- promotes memory space despite of the processing time. The modified C4.5 algorithm requires three entries tables: a target table, a neighbor table, and a spatial index join that contains the possible spatial relationship among the objects in the target table and those in the neighbor table. Thus, the proposed algorithms are applied to a spatial data pattern in the accidentology domain. A comparative study of our approach with other works of classification by spatial decision trees will be detailed.

Keywords: C4.5 algorithm, decision trees, S-CART, spatial data mining

Procedia PDF Downloads 615

Authors: Houman Tamaddon, Mehrdad Behnia, Masud Behnia

Abstract:

An accurate study of blood flow is associated with an accurate vascular pattern and geometrical properties of the organ of interest. Due to the complexity of vascular networks and poor accessibility in vivo, it is challenging to reconstruct the entire vasculature of any organ experimentally. The objective of this study is to introduce an innovative approach for the reconstruction of a full vascular tree from available morphometric data. Our method consists of implementing morphometric data on those parts of the vascular tree that are smaller than the resolution of medical imaging methods. This technique reconstructs the entire arterial tree down to the capillaries. Vessels greater than 2 mm are obtained from direct volume and surface analysis using contrast enhanced computed tomography (CT). Vessels smaller than 2mm are reconstructed from available morphometric and distensibility data and rearranged by applying Murray’s Laws. Implementation of morphometric data to reconstruct the branching pattern and applying Murray’s Laws to every vessel bifurcation simultaneously, lead to an accurate vascular tree reconstruction. The reconstruction algorithm generates full arterial tree topography down to the first capillary bifurcation. Geometry of each order of the vascular tree is generated separately to minimize the construction and simulation time. The node-to-node connectivity along with the diameter and length of every vessel segment is established and order numbers, according to the diameter-defined Strahler system, are assigned. During the simulation, we used the averaged flow rate for each order to predict the pressure drop and once the pressure drop is predicted, the flow rate is corrected to match the computed pressure drop for each vessel. The final results for 3 cardiac cycles is presented and compared to the clinical data.

Keywords: blood flow, morphometric data, vascular tree, Strahler ordering system

Procedia PDF Downloads 274

Authors: A. S. Cheliyan, S. K. Bhattacharyya

Abstract:

Dynamic Positioning System (DPS) is employed in marine vessels of the offshore oil and gas industry. It is a computer controlled system to automatically maintain a ship’s position and heading by using its own thrusters. Reliability assessment of the same can be analyzed through conventional fault tree. However, the complex behaviour like sequence failure, redundancy management and priority of failing of events cannot be analyzed by the conventional fault trees. The Dynamic Fault Tree (DFT) addresses these shortcomings of conventional Fault Tree by defining additional gates called dynamic gates. Monte Carlo based simulation approach has been adopted for the dynamic gates. This method of realistic modeling of DPS gives meaningful insight into the system reliability and the ability to improve the same.

Keywords: dynamic positioning system, dynamic fault tree, Monte Carlo simulation, reliability assessment

Procedia PDF Downloads 775

Authors: Qiang Zhang, Chun Yuan

Abstract:

Since the PU (Prediction Unit) decision process is the most time consuming part of the emerging HEVC (High Efficient Video Coding) standardin intra and inter frame coding, this paper proposes the fast PU decision algorithm and speed up the algorithm using CUDA (Compute Unified Device Architecture). In intra frame coding, the fast PU decision algorithm uses the texture features to skip intra-frame prediction or terminal the intra-frame prediction for smaller PU size. In inter frame coding of HEVC, the fast PU decision algorithm takes use of the similarity of its own two Nx2N size PU's motion vectors and the hierarchical structure of CU (Coding Unit) partition to skip some modes of PU partition, so as to reduce the motion estimation times. The accelerate algorithm using CUDA is based on the fast PU decision algorithm which uses the GPU to make the motion search and the gradient computation could be parallel computed. The proposed algorithm achieves up to 57% time saving compared to the HM 10.0 with little rate-distortion losses (0.043dB drop and 1.82% bitrate increase on average).

Keywords: HEVC, PU decision, inter prediction, intra prediction, CUDA, parallel

Procedia PDF Downloads 399

Authors: Kemal Polat

Abstract:

In this study, it has been estimated whether there is skin by using RBG values obtained from the camera and k-nearest neighbor (k-NN) classifier. The dataset used in this study has an unbalanced distribution and a linearly non-separable structure. This problem can also be called a big data problem. The Skin dataset was taken from UCI machine learning repository. As the classifier, we have used the k-NN method to handle this big data problem. For k value of k-NN classifier, we have used as 1. To train and test the k-NN classifier, 50-50% training-testing partition has been used. As the performance metrics, TP rate, FP Rate, Precision, recall, f-measure and AUC values have been used to evaluate the performance of k-NN classifier. These obtained results are as follows: 0.999, 0.001, 0.999, 0.999, 0.999, and 1,00. As can be seen from the obtained results, this proposed method could be used to predict whether the image is skin or not.

Keywords: k-NN classifier, skin or non-skin classification, RGB values, classification

Procedia PDF Downloads 249

Authors: Francesca Schiavello

Abstract:

This work aims to create a hybrid algorithm, combining Quantum Approximate Optimization Algorithm (QAOA) with an Evolutionary Algorithm (EA) in the place of traditional gradient based optimization processes. QAOA’s were first introduced in 2014, where, at the time, their algorithm performed better than the traditional best known classical algorithm for Max-cut graphs. Whilst classical algorithms have improved since then and have returned to being faster and more efficient, this was a huge milestone for quantum computing, and their work is often used as a benchmarking tool and a foundational tool to explore variants of QAOA’s. This, alongside with other famous algorithms like Grover’s or Shor’s, highlights to the world the potential that quantum computing holds. It also presents the reality of a real quantum advantage where, if the hardware continues to improve, this could constitute a revolutionary era. Given that the hardware is not there yet, many scientists are working on the software side of things in the hopes of future progress. Some of the major limitations holding back quantum computing are the quality of qubits and the noisy interference they generate in creating solutions, the barren plateaus that effectively hinder the optimization search in the latent space, and the availability of number of qubits limiting the scale of the problem that can be solved. These three issues are intertwined and are part of the motivation for using EAs in this work. Firstly, EAs are not based on gradient or linear optimization methods for the search in the latent space, and because of their freedom from gradients, they should suffer less from barren plateaus. Secondly, given that this algorithm performs a search in the solution space through a population of solutions, it can also be parallelized to speed up the search and optimization problem. The evaluation of the cost function, like in many other algorithms, is notoriously slow, and the ability to parallelize it can drastically improve the competitiveness of QAOA’s with respect to purely classical algorithms. Thirdly, because of the nature and structure of EA’s, solutions can be carried forward in time, making them more robust to noise and uncertainty. Preliminary results show that the EA algorithm attached to QAOA can perform on par with the traditional QAOA with a Cobyla optimizer, which is a linear based method, and in some instances, it can even create a better Max-Cut. Whilst the final objective of the work is to create an algorithm that can consistently beat the original QAOA, or its variants, due to either speedups or quality of the solution, this initial result is promising and show the potential of EAs in this field. Further tests need to be performed on an array of different graphs with the parallelization aspect of the work commencing in October 2023 and tests on real hardware scheduled for early 2024.

Keywords: evolutionary algorithm, max cut, parallel simulation, quantum optimization

Procedia PDF Downloads 60

Authors: C. Rekha, G. N. Krishnamurthy

Abstract:

This paper attempts to develop a different approach for key scheduling algorithm which uses both Blowfish and AES algorithms. The main drawback of Blowfish algorithm is, it takes more time to create the S-box entries. To overcome this, we are replacing process of S-box creation in blowfish, by using key dependent S-box creation from AES without affecting the basic operation of blowfish. The method proposed in this paper uses good features of blowfish as well as AES and also this paper demonstrates the performance of blowfish and new algorithm by considering different aspects of security namely Encryption Quality, Key Sensitivity, and Correlation of horizontally adjacent pixels in an encrypted image.

Keywords: AES, blowfish, correlation coefficient, encryption quality, key sensitivity, s-box

Procedia PDF Downloads 226

Authors: R. Bouhennache, T. Bouden, A. A. Taleb, A. Chaddad

Abstract:

In this paper we propose a method to map the urban areas. The method uses an arithmetic calculation processed from the land features indexes and Tasseled cap transformation TC of multi spectral Thematic Mapper Landsat TM image. For this purpose the derived indexes image from the original image such SAVI the soil adjusted vegetation index, UI the urban Index, and EBBI the enhanced built up and bareness index were staked to form a new image and the bands were uncorrelated, also the Spectral Angle Mapper (SAM) and Spectral Information Divergence (SID) supervised classification approaches were first applied on the new image TM data using the reference spectra of the spectral library and subsequently the four urban, vegetation, water and soil land cover categories were extracted with their accuracy assessment.The urban features were represented using a logic calculation applied to the brightness, UI-SAVI, NDBI-greenness and EBBI- brightness data sets. The study applied to Blida and mentioned that the urban features can be mapped with an accuracy ranging from 92 % to 95%.

Keywords: EBBI, SAVI, Tasseled Cap Transformation, UI

Procedia PDF Downloads 484

Authors: Vahid Anari, Leila Shahmohammadi

Abstract:

Diagnosing and interpreting manually from a large cohort dataset of immunohistochemically stained tissue of tumors using an optical microscope involves subjectivity and also is tedious for pathologist specialists. Moreover, digital pathology today represents more of an evolution than a revolution in pathology. In this paper, we develop and test an unsupervised algorithm that can automatically enhance the IHC image of a meningioma tumor and classify cells into positive (proliferative) and negative (normal) cells. A dataset including 150 images is used to test the scheme. In addition, a new adaptive color image enhancement method is proposed based on a vector directional filter (VDF) and statistical properties of filtering the window. Since the cells are distinguishable by the human eye, the accuracy and stability of the algorithm are quantitatively compared through application to a wide variety of real images.

Keywords: digital pathology, cell segmentation, immunohistochemically, noise reduction

Procedia PDF Downloads 67

Authors: Chhabi Nigam, S. Ramakrishnan

Abstract:

This paper brings out the analysis of the airborne Synthetic Aperture Radar (SAR) data to study the impact of Doppler centroid on Image quality and geocoding accuracy from the perspective of Stripmap mode of data acquisition. Although in Stripmap mode of data acquisition radar beam points at 90 degrees broad side (side looking), shift in the Doppler centroid is invariable due to platform motion. In-accurate estimation of Doppler centroid leads to poor image quality and image miss-registration. The effect of Doppler centroid is analyzed in this paper using multiple sets of data collected from airborne platform. Occurrences of ghost (ambiguous) targets and their power levels have been analyzed that impacts appropriate choice of PRF. Effect of aircraft attitudes (roll, pitch and yaw) on the Doppler centroid is also analyzed with the collected data sets. Various stages of the RDA (Range Doppler Algorithm) algorithm used for image formation in Stripmap mode, range compression, Doppler centroid estimation, azimuth compression, range cell migration correction are analyzed to find the performance limits and the dependence of the imaging geometry on the final image. The ability of Doppler centroid estimation to enhance the imaging accuracy for registration are also illustrated in this paper. The paper also tries to bring out the processing of low squint SAR data, the challenges and the performance limits imposed by the imaging geometry and the platform dynamics on the final image quality metrics. Finally, the effect on various terrain types, including land, water and bright scatters is also presented.

Keywords: ambiguous target, Doppler Centroid, image registration, Airborne SAR

Procedia PDF Downloads 218

Authors: N. Ghatwary, A. Ahmed, H. Jalab

Abstract:

In this paper, an approach for the liver tumor detection in computed tomography (CT) images is represented. The detection process is based on classifying the features of target liver cell to either tumor or non-tumor. Fractional differential (FD) is applied for enhancement of Liver CT images, with the aim of enhancing texture and edge features. Later on, a fusion method is applied to merge between the various enhanced images and produce a variety of feature improvement, which will increase the accuracy of classification. Each image is divided into NxN non-overlapping blocks, to extract the desired features. Support vector machines (SVM) classifier is trained later on a supplied dataset different from the tested one. Finally, the block cells are identified whether they are classified as tumor or not. Our approach is validated on a group of patients’ CT liver tumor datasets. The experiment results demonstrated the efficiency of detection in the proposed technique.

Keywords: fractional differential (FD), computed tomography (CT), fusion, aplha, texture features.

Procedia PDF Downloads 359

Authors: Hankyu Lee, Seoung Bum Kim

Abstract:

The issue with imbalance and overlapping in the class distribution becomes important in various applications of data mining. The imbalanced dataset is a special case in classification problems in which the number of observations of one class (i.e., major class) heavily exceeds the number of observations of the other class (i.e., minor class). Overlapped dataset is the case where many observations are shared together between the two classes. Imbalanced and overlapped data can be frequently found in many real examples including fraud and abuse patients in healthcare, quality prediction in manufacturing, text classification, oil spill detection, remote sensing, and so on. The class imbalance and overlap problem is the challenging issue because this situation degrades the performance of most of the standard classification algorithms. In this study, we propose a classification procedure that can effectively handle imbalanced and overlapped datasets by splitting data space into three parts: nonoverlapping, light overlapping, and severe overlapping and applying the classification algorithm in each part. These three parts were determined based on the Hausdorff distance and the margin of the modified support vector machine. An experiments study was conducted to examine the properties of the proposed method and compared it with other classification algorithms. The results showed that the proposed method outperformed the competitors under various imbalanced and overlapped situations. Moreover, the applicability of the proposed method was demonstrated through the experiment with real data.

Keywords: classification, imbalanced data with class overlap, split data space, support vector machine

Procedia PDF Downloads 308

Authors: Michael Aupetit, Mahmoud Al-ismail, Khaled Mohamed

Abstract:

Cancer is a major public health problem all over the world. Breast cancer has the highest incidence rate over all cancers for women in Qatar making its study a top priority of the country. Human cancer is a dynamic disease that develops over an extended period through the accumulation of a series of genetic alterations. A Darwinian process drives the tumor cells toward higher malignancy growing the branches of a progression tree in the space of genes expression. Although it is not possible to track these genetic alterations dynamically for one patient, it is possible to reconstruct the progression tree from the aggregation of thousands of tumor cells’ genetic profiles from thousands of different patients at different stages of the disease. Analyzing the progression tree is a way to detect pivotal molecular events that drive the malignant evolution and to provide a guide for the development of cancer diagnostics, prognostics and targeted therapeutics. In this work we present the development of a Visual Analytic web platform CanVis enabling users to upload gene-expression data and analyze their progression tree. The server computes the progression tree based on state-of-the-art techniques and allows an interactive visual exploration of this tree and the gene-expression data along its branching structure helping to discover potential driver genes.

Keywords: breast cancer, progression tree, visual analytics, web platform

Procedia PDF Downloads 419

Authors: Tigabu Dagne Akal

Abstract:

While advances in computer and communications technology have made the network ubiquitous, they have also rendered networked systems vulnerable to malicious attacks devised from a distance. These attacks or intrusions start with attackers infiltrating a network through a vulnerable host and then launching further attacks on the local network or Intranet. Nowadays, system administrators and network professionals can attempt to prevent such attacks by developing intrusion detection tools and systems using data mining technology. In this study, the experiments were conducted following the Knowledge Discovery in Database Process Model. The Knowledge Discovery in Database Process Model starts from selection of the datasets. The dataset used in this study has been taken from Massachusetts Institute of Technology Lincoln Laboratory. After taking the data, it has been pre-processed. The major pre-processing activities include fill in missed values, remove outliers; resolve inconsistencies, integration of data that contains both labelled and unlabelled datasets, dimensionality reduction, size reduction and data transformation activity like discretization tasks were done for this study. A total of 21,533 intrusion records are used for training the models. For validating the performance of the selected model a separate 3,397 records are used as a testing set. For building a predictive model for intrusion detection J48 decision tree and the Naïve Bayes algorithms have been tested as a classification approach for both with and without feature selection approaches. The model that was created using 10-fold cross validation using the J48 decision tree algorithm with the default parameter values showed the best classification accuracy. The model has a prediction accuracy of 96.11% on the training datasets and 93.2% on the test dataset to classify the new instances as normal, DOS, U2R, R2L and probe classes. The findings of this study have shown that the data mining methods generates interesting rules that are crucial for intrusion detection and prevention in the networking industry. Future research directions are forwarded to come up an applicable system in the area of the study.

Keywords: intrusion detection, data mining, computer science, data mining

Procedia PDF Downloads 297

Authors: Shannon Steinmetz, Ellen Gethner

Abstract:

In this paper, a theoretical foundation is developed for patterned segmentation of audio using the geometry of music and statistical manifold. We demonstrate image content clustering using conic space sonification. The algorithm takes a geodesic curve as a model estimator of the three-parameter Gamma distribution. The random variable is parameterized by musical centricity and centric velocity. Model parameters predict audio segmentation in the form of duration and frame count based on the likelihood of musical geometry transition. We provide an example using a database of randomly selected images, resulting in statistically significant clusters of similar image content.

Keywords: sonification, musical information geometry, image, content extraction, automated quantification, audio segmentation, pattern recognition

Procedia PDF Downloads 240

Authors: G. R. Navaneesh, E. Nagarajan, C. H. Rajam Raju

Abstract:

Nowadays the multimedia data are used to store some secure information. All previous methods allocate a space in image for data embedding purpose after encryption. In this paper, we propose a novel method by reserving space in image with a boundary surrounded before encryption with a traditional RDH algorithm, which makes it easy for the data hider to reversibly embed data in the encrypted images. The proposed method can achieve real time performance, that is, data extraction and image recovery are free of any error. A secure transmission process is also discussed in this paper, which improves the efficiency by ten times compared to other processes as discussed.

Keywords: secure communication, reserving room before encryption, least significant bits, image encryption, reversible data hiding

Procedia PDF Downloads 413

Authors: Sukhleen Kaur

Abstract:

In security groundwork, Intrusion Detection System (IDS) has become an important component. The IDS has received increasing attention in recent years. IDS is one of the effective way to detect different kinds of attacks and malicious codes in a network and help us to secure the network. Data mining techniques can be implemented to IDS, which analyses the large amount of data and gives better results. Data mining can contribute to improving intrusion detection by adding a level of focus to anomaly detection. So far the study has been carried out on finding the attacks but this paper detects the malicious files. Some intruders do not attack directly, but they hide some harmful code inside the files or may corrupt those file and attack the system. These files are detected according to some defined parameters which will form two lists of files as normal files and harmful files. After that data mining will be performed. In this paper a hybrid classifier has been used via Naive Bayes and Ripper classification methods. The results show how the uploaded file in the database will be tested against the parameters and then it is characterised as either normal or harmful file and after that the mining is performed. Moreover, when a user tries to mine on harmful file it will generate an exception that mining cannot be made on corrupted or harmful files.

Keywords: data mining, association, classification, clustering, decision tree, intrusion detection system, misuse detection, anomaly detection, naive Bayes, ripper

Procedia PDF Downloads 414

Authors: Rikson Gultom

Abstract:

Hate Speech and Abusive language on social media is difficult to detect, usually, it is detected after it becomes viral in cyberspace, of course, it is too late for prevention. An early detection system that has a fairly good accuracy is needed so that it can reduce conflicts that occur in society caused by postings on social media that attack individuals, groups, and governments in Indonesia. The purpose of this study is to find an early detection model on Twitter social media using machine learning that has high accuracy from several machine learning methods studied. In this study, the support vector machine (SVM), Naïve Bayes (NB), and Random Forest Decision Tree (RFDT) methods were compared with the Support Vector machine with genetic algorithm (SVM-GA), Nave Bayes with genetic algorithm (NB-GA), and Random Forest Decision Tree with Genetic Algorithm (RFDT-GA). The study produced a comparison table for the accuracy of the hate speech and abusive language detection model, and presented it in the form of a graph of the accuracy of the six algorithms developed based on the Indonesian-language Twitter dataset, and concluded the best model with the highest accuracy.

Keywords: abusive language, hate speech, machine learning, optimization, social media

Procedia PDF Downloads 129

Authors: Yavuz Unal, Kemal Polat, H. Erdinc Kocer

Abstract:

In this study, a decision support system comprising three stages has been proposed to detect the disc abnormalities of the lumbar region. In the first stage named the feature extraction, T2-weighted sagittal and axial Magnetic Resonance Images (MRI) were taken from 55 people and then 27 appearance and shape features were acquired from both sagittal and transverse images. In the second stage named the feature weighting process, k-means clustering based feature weighting (KMCBFW) proposed by Gunes et al. Finally, in the third stage named the classification process, the classifier algorithms including multi-layer perceptron (MLP- neural network), support vector machine (SVM), Naïve Bayes, and decision tree have been used to classify whether the subject has lumbar disc or not. In order to test the performance of the proposed method, the classification accuracy (%), sensitivity, specificity, precision, recall, f-measure, kappa value, and computation times have been used. The best hybrid model is the combination of k-means clustering based feature weighting and decision tree in the detecting of lumbar disc disease based on both sagittal and axial MR images.

Keywords: lumbar disc abnormality, lumbar MRI, lumbar spine, hybrid models, hybrid features, k-means clustering based feature weighting

Procedia PDF Downloads 521

Authors: Jheanel E. Estrada, Larry A. Vea

Abstract:

This study developed model to detect proper/improper sitting posture using the built in web camera which detects the upper body points’ location and distances (chin, manubrium and acromion process). It also established relationships of human body frames and proper sitting posture. The models were developed by training some well-known classifiers such as KNN, SVM, MLP, and Decision Tree using the data collected from 60 students of different body frames. Decision Tree classifier demonstrated the most promising model performance with an accuracy of 95.35% and a kappa of 0.907 for head and shoulder posture. Results also showed that there were relationships between body frame and posture through Body Mass Index.

Keywords: posture, spinal points, gyroscope, image processing, ergonomics

Procedia PDF Downloads 329

Authors: Mahasa Zahirnia

Abstract:

The purpose of this paper is to explore new and effective ways of optimizing satellite images using artificial intelligence, and the process of implementing reinforcement learning to enhance the quality of data captured within the image. In our implementation of Bellman's Reinforcement Learning equations, associated state diagrams, and multi-stage image processing, we were able to enhance image quality, detect and define objects. Reinforcement learning is the differentiator in the area of artificial intelligence, and Q-Learning relies on trial and error to achieve its goals. The reward system that is embedded in Q-Learning allows the agent to self-evaluate its performance and decide on the best possible course of action based on the current and future environment. Results show that within a simulated environment, built on the images that are commercially available, the rate of detection was 40-90%. Reinforcement learning through Q-Learning algorithm is not just desired but required design criteria for image optimization and enhancements. The proposed methods presented are a cost effective method of resolving uncertainty of the data because reinforcement learning finds ideal policies to manage the process using a smaller sample of images.

Keywords: Q-learning, image optimization, reinforcement learning, Markov decision process

Procedia PDF Downloads 216

Authors: Majid Khalili

Abstract:

This paper deals with a bi-objective hybrid no-wait flowshop scheduling problem minimizing the makespan and total weighted tardiness, in which we consider transportation times between stages. Obtaining an optimal solution for this type of complex, large-sized problem in reasonable computational time by using traditional approaches and optimization tools is extremely difficult. This paper presents a new multi-objective variable neighborhood algorithm (MOVNS). A set of experimental instances are carried out to evaluate the algorithm by advanced multi-objective performance measures. The algorithm is carefully evaluated for its performance against available algorithm by means of multi-objective performance measures and statistical tools. The related results show that a variant of our proposed MOVNS provides sound performance comparing with other algorithms.

Keywords: no-wait hybrid flowshop scheduling; multi-objective variable neighborhood algorithm; makespan; total weighted tardiness

Procedia PDF Downloads 418

Authors: Xiuying Wang, Changliang Li, Bo Xu

Abstract:

A fixed-length feature vector is required for many machine learning algorithms in NLP field. Word embeddings have been very successful at learning lexical information. However, they cannot capture the compositional meaning of sentences, which prevents them from a deeper understanding of language. In this paper, we introduce a novel hierarchical tree long short-term memory (HTLSTM) model that learns vector representations for sentences of arbitrary syntactic type and length. We propose to split one sentence into three hierarchies: short phrase, long phrase and full sentence level. The HTLSTM model gives our algorithm the potential to fully consider the hierarchical information and long-term dependencies of language. We design the experiments on both English and Chinese corpus to evaluate our model on sentiment analysis task. And the results show that our model outperforms several existing state of the art approaches significantly.

Keywords: deep learning, hierarchical tree long short-term memory, sentence representation, sentiment analysis

Procedia PDF Downloads 349

Authors: Mohamed I. Mahrous, Mohamed G. Ashmawy

Abstract:

Hybrid genetic algorithm (HGA) is proposed in this paper to determine the economic scheduling of electric power generation over a fixed time period under various system and operational constraints. The proposed technique can outperform conventional genetic algorithms (CGAs) in the sense that HGA make it possible to improve both the quality of the solution and reduce the computing expenses. In contrast, any carefully designed GA is only able to balance the exploration and the exploitation of the search effort, which means that an increase in the accuracy of a solution can only occure at the sacrifice of convergent speed, and vice visa. It is unlikely that both of them can be improved simultaneously. The proposed hybrid scheme is developed in such a way that a simple GA is acting as a base level search, which makes a quick decision to direct the search towards the optimal region, and a local search method (pattern search technique) is next employed to do the fine tuning. The aim of the strategy is to achieve the cost reduction within a reasonable computing time. The effectiveness of the proposed hybrid technique is verified on two real public electricity supply systems with 13 and 40 generator units respectively. The simulation results obtained with the HGA for the two real systems are very encouraging with regard to the computational expenses and the cost reduction of power generation.

Keywords: genetic algorithms, economic dispatch, pattern search

Procedia PDF Downloads 445

Authors: Thomas Bryan, Veton Kepuska, Ivica Kostanic

Abstract:

Speaker recognition is performed in high Additive White Gaussian Noise (AWGN) environments using principals of Computational Auditory Scene Analysis (CASA). CASA methods often classify sounds from images in the time-frequency (T-F) plane using spectrograms or cochleargrams as the image. In this paper atomic decomposition implemented by matching pursuit performs a transform from time series speech signals to the T-F plane. The atomic decomposition creates a sparsely populated T-F vector in “weight space” where each populated T-F position contains an amplitude weight. The weight space vector along with the atomic dictionary represents a denoised, compressed version of the original signal. The arraignment or of the atomic indices in the T-F vector are used for classification. Unsupervised feature learning implemented by a sparse autoencoder learns a single dictionary of basis features from a collection of envelope samples from all speakers. The approach is demonstrated using pairs of speakers from the TIMIT data set. Pairs of speakers are selected randomly from a single district. Each speak has 10 sentences. Two are used for training and 8 for testing. Atomic index probabilities are created for each training sentence and also for each test sentence. Classification is performed by finding the lowest Euclidean distance between then probabilities from the training sentences and the test sentences. Training is done at a 30dB Signal-to-Noise Ratio (SNR). Testing is performed at SNR’s of 0 dB, 5 dB, 10 dB and 30dB. The algorithm has a baseline classification accuracy of ~93% averaged over 10 pairs of speakers from the TIMIT data set. The baseline accuracy is attributable to short sequences of training and test data as well as the overall simplicity of the classification algorithm. The accuracy is not affected by AWGN and produces ~93% accuracy at 0dB SNR.

Keywords: time-frequency plane, atomic decomposition, envelope sampling, Gabor atoms, matching pursuit, sparse dictionary learning, sparse autoencoder

Procedia PDF Downloads 290

Authors: Levent Basayigit, Sinan Demir, Yusuf Ucar, Burhan Kara

Abstract:

Many research projects require accurate delineation of the different land cover type of the agricultural area. Especially it is critically important for the definition of specific plants like cannabis. However, the complexity of vegetation stands structure, abundant vegetation species, and the smooth transition between different seconder section stages make vegetation classification difficult when using traditional approaches such as the maximum likelihood classifier. Most of the time, classification distinguishes only between trees/annual or grain. It has been difficult to accurately determine the cannabis mixed with other plants. In this paper, a mixed distribution models approach is applied to classify pure and mix cannabis parcels using Worldview-2 imagery in the Lakes region of Turkey. Five different land use types (i.e. sunflower, maize, bare soil, and cannabis) were identified in the image. A constrained Gaussian mixture discriminant analysis (GMDA) was used to unmix the image. In the study, 255 reflectance ratios derived from spectral signatures of seven bands (Blue-Green-Yellow-Red-Rededge-NIR1-NIR2) were randomly arranged as 80% for training and 20% for test data. Gaussian mixed distribution model approach is proved to be an effective and convenient way to combine very high spatial resolution imagery for distinguishing cannabis vegetation. Based on the overall accuracies of the classification, the Gaussian mixed distribution model was found to be very successful to achieve image classification tasks. This approach is sensitive to capture the illegal cannabis planting areas in the large plain. This approach can also be used for monitoring and determination with spectral reflections in illegal cannabis planting areas.

Keywords: Gaussian mixture discriminant analysis, spectral mixture model, Worldview-2, land parcels

Procedia PDF Downloads 197

Authors: Jonathan Sohn

Abstract:

Deep neural networks (DNNs) have shown state-of-the-art performance for many applications, including computer vision, natural language processing, and speech recognition. Recently, adversarial attacks have been studied in the context of deep neural networks, which aim to alter the results of deep neural networks by modifying the inputs slightly. For example, an adversarial attack on a DNN used for object detection can cause the DNN to miss certain objects. As a result, the reliability of DNNs is undermined by their lack of robustness against adversarial attacks, raising concerns about their use in safety-critical applications such as autonomous driving. In this paper, we focus on studying the adversarial attacks and defenses on DNNs for image classification. There are two types of adversarial attacks studied which are fast gradient sign method (FGSM) attack and projected gradient descent (PGD) attack. A DNN forms decision boundaries that separate the input images into different categories. The adversarial attack slightly alters the image to move over the decision boundary, causing the DNN to misclassify the image. FGSM attack obtains the gradient with respect to the image and updates the image once based on the gradients to cross the decision boundary. PGD attack, instead of taking one big step, repeatedly modifies the input image with multiple small steps. There is also another type of attack called the target attack. This adversarial attack is designed to make the machine classify an image to a class chosen by the attacker. We can defend against adversarial attacks by incorporating adversarial examples in training. Specifically, instead of training the neural network with clean examples, we can explicitly let the neural network learn from the adversarial examples. In our experiments, the digit recognition accuracy on the MNIST dataset drops from 97.81% to 39.50% and 34.01% when the DNN is attacked by FGSM and PGD attacks, respectively. If we utilize FGSM training as a defense method, the classification accuracy greatly improves from 39.50% to 92.31% for FGSM attacks and from 34.01% to 75.63% for PGD attacks. To further improve the classification accuracy under adversarial attacks, we can also use a stronger PGD training method. PGD training improves the accuracy by 2.7% under FGSM attacks and 18.4% under PGD attacks over FGSM training. It is worth mentioning that both FGSM and PGD training do not affect the accuracy of clean images. In summary, we find that PGD attacks can greatly degrade the performance of DNNs, and PGD training is a very effective way to defend against such attacks. PGD attacks and defence are overall significantly more effective than FGSM methods.

Keywords: deep neural network, adversarial attack, adversarial defense, adversarial machine learning

Procedia PDF Downloads 196

Authors: Kashika P. H.

Abstract:

The development of computer vision systems has been greatly aided by the efficient and precise detection of images and videos. Although the ability to recognize and comprehend images is a strength of the human brain, employing technology to tackle this issue is exceedingly challenging. In the past few years, the use of Deep Learning algorithms to treat object detection has dramatically expanded. One of the key issues in the realm of image recognition is the recognition and detection of certain notable people from randomly acquired photographs. Face recognition uses a way to identify, assess, and compare faces for a variety of purposes, including user identification, user counting, and classification. With the aid of an accessible deep learning-based API, this article intends to recognize various faces of people and their facial descriptors more accurately. The purpose of this study is to locate suitable individuals and deliver accurate information about them by using the Amazon Rekognition system to identify a specific human from a vast image dataset. We have chosen the Amazon Rekognition system, which allows for more accurate face analysis, face comparison, and face search, to tackle this difficulty.

Keywords: Amazon rekognition, API, deep learning, computer vision, face detection, text detection

Procedia PDF Downloads 107

Authors: A. Biran, P. Sobhe Bidari, A. Almazroe, V. Lakshminarayanan, K. Raahemifar

Abstract:

Diabetic Retinopathy (DR) is a severe retinal disease which is caused by diabetes mellitus. It leads to blindness when it progress to proliferative level. Early indications of DR are the appearance of microaneurysms, hemorrhages and hard exudates. In this paper, an automatic algorithm for detection of DR has been proposed. The algorithm is based on combination of several image processing techniques including Circular Hough Transform (CHT), Contrast Limited Adaptive Histogram Equalization (CLAHE), Gabor filter and thresholding. Also, Support Vector Machine (SVM) Classifier is used to classify retinal images to normal or abnormal cases including non-proliferative or proliferative DR. The proposed method has been tested on images selected from Structured Analysis of the Retinal (STARE) database using MATLAB code. The method is perfectly able to detect DR. The sensitivity specificity and accuracy of this approach are 90%, 87.5%, and 91.4% respectively.

Keywords: diabetic retinopathy, fundus images, STARE, Gabor filter, support vector machine

Procedia PDF Downloads 294