Browsing by Author "Gabbouj, Moncef"

Now showing 1 - 20 of 53

Citation - WoS: 238
1-D Convolutional Neural Networks for Signal Processing Applications
(IEEE, 2019) Kiranyaz, Serkan; İnce, Türker; Abdeljaber, Osama; Avcı, Onur; Gabbouj, Moncef
1D Convolutional Neural Networks (CNNs) have recently become the state-of-the-art technique for crucial signal processing applications such as patient-specific ECG classification, structural health monitoring, anomaly detection in power electronics circuitry and motor-fault detection. This is an expected outcome as there are numerous advantages of using an adaptive and compact 1D CNN instead of a conventional (2D) deep counterparts. First of all, compact 1D CNNs can be efficiently trained with a limited dataset of 1D signals while the 2D deep CNNs, besides requiring 1D to 2D data transformation, usually need datasets with massive size, e.g., in the Big Data scale in order to prevent the well-known overfitting problem. 1D CNNs can directly be applied to the raw signal (e.g., current, voltage, vibration, etc.) without requiring any pre-or postprocessing such as feature extraction, selection, dimension reduction, denoising, etc. Furthermore, due to the simple and compact configuration of such adaptive 1D CNNs that perform only linear 1D convolutions (scalar multiplications and additions), a real-time and low-cost hardware implementation is feasible. This paper reviews the major signal processing applications of compact 1D CNNs with a brief theoretical background. We will present their state-of-the-art performances and conclude with focusing on some major properties.
Citation - WoS: 1893
Citation - Scopus: 2230
1d Convolutional Neural Networks and Applications: a Survey
(Academic Press Ltd- Elsevier Science Ltd, 2021) Kiranyaz, Serkan; Avcı, Onur; Abdeljaber, Osama; İnce, Türker; Gabbouj, Moncef; Inman, Daniel J.
During the last decade, Convolutional Neural Networks (CNNs) have become the de facto standard for various Computer Vision and Machine Learning operations. CNNs are feed-forward Artificial Neural Networks (ANNs) with alternating convolutional and subsampling layers. Deep 2D CNNs with many hidden layers and millions of parameters have the ability to learn complex objects and patterns providing that they can be trained on a massive size visual database with ground-truth labels. With a proper training, this unique ability makes them the primary tool for various engineering applications for 2D signals such as images and video frames. Yet, this may not be a viable option in numerous applications over 1D signals especially when the training data is scarce or application specific. To address this issue, 1D CNNs have recently been proposed and immediately achieved the state-of-the-art performance levels in several applications such as personalized biomedical data classification and early diagnosis, structural health monitoring, anomaly detection and identification in power electronics and electrical motor fault detection. Another major advantage is that a real-time and low-cost hardware implementation is feasible due to the simple and compact configuration of 1D CNNs that perform only 1D convolutions (scalar multiplications and additions). This paper presents a comprehensive review of the general architecture and principals of 1D CNNs along with their major engineering applications, especially focused on the recent progress in this field. Their state-of-the-art performance is highlighted concluding with their unique properties. The benchmark datasets and the principal 1D CNN software used in those applications are also publicly shared in a dedicated website. While there has not been a paper on the review of 1D CNNs and its applications in the literature, this paper fulfills this gap. (C) 2020 The Author(s). Published by Elsevier Ltd.
Citation - WoS: 112
Citation - Scopus: 130
Analysis of High-Dimensional Phase Space Via Poincare Section for Patient-Specific Seizure Detection
(IEEE-Inst Electrical Electronics Engineers Inc, 2016) Zabihi, Morteza; Kiranyaz, Serkan; Rad, Ali Bahrami; Katsaggelos, Aggelos K.; Gabbouj, Moncef; İnce, Türker
In this paper, the performance of the phase space representation in interpreting the underlying dynamics of epileptic seizures is investigated and a novel patient-specific seizure detection approach is proposed based on the dynamics of EEG signals. To accomplish this, the trajectories of seizure and nonseizure segments are reconstructed in a high dimensional space using time-delay embedding method. Afterwards, Principal Component Analysis (PCA) was used in order to reduce the dimension of the reconstructed phase spaces. The geometry of the trajectories in the lower dimensions is then characterized using Poincare section and seven features were extracted from the obtained intersection sequence. Once the features are formed, they are fed into a two-layer classification scheme, comprising the Linear Discriminant Analysis (LDA) and Naive Bayesian classifiers. The performance of the proposed method is then evaluated over the CHB-MIT benchmark database and the proposed approach achieved 88.27% sensitivity and 93.21% specificity on average with 25% training data. Finally, we perform comparative performance evaluations against the state-of-the-art methods in this domain which demonstrate the superiority of the proposed method.
Citation - Scopus: 3
Audio-Based Anomaly Detection in Industrial Machines Using Deep One-Class Support Vector Data Description
(IEEE, 2025) Kilickaya, Sertac; Ahishali, Mete; Celebioglu, Cansu; Sohrab, Fahad; Eren, Levent; Ince, Turker; Gabbouj, Moncef
The frequent breakdowns and malfunctions of industrial equipment have driven increasing interest in utilizing cost-effective and easy-to-deploy sensors, such as microphones, for effective condition monitoring of machinery. Microphones offer a low-cost alternative to widely used condition monitoring sensors with their high bandwidth and capability to detect subtle anomalies that other sensors might have less sensitivity. In this study, we investigate malfunctioning industrial machines to evaluate and compare anomaly detection performance across different machine types and fault conditions. Log-Mel spectrograms of machinery sound are used as input, and the performance is evaluated using the area under the curve (AUC) score for two different methods: baseline dense autoencoder (AE) and oneclass deep Support Vector Data Description (deep SVDD) with different subspace dimensions. Our results over the MIMII sound dataset demonstrate that the deep SVDD method with a subspace dimension of 2 provides superior anomaly detection performance, achieving average AUC scores of 0.84, 0.80, and 0.69 for 6 dB, 0 dB, and -6 dB signal-to-noise ratios (SNRs), respectively, compared to 0.82, 0.72, and 0.64 for the baseline model. Moreover, deep SVDD requires 7.4 times fewer trainable parameters than the baseline dense AE, emphasizing its advantage in both effectiveness and computational efficiency.
Citation - WoS: 17
Citation - Scopus: 21
Automated Patient-Specific Classification of Premature Ventricular Contractions
(IEEE, 2008) İnce, Türker; Kiranyaz, Serkan; Gabbouj, Moncef
In this paper, we present an automated patient-specific electrocardiogram (ECG) beat classifier designed or accurate defection of premature ventricular contractions (PVCs). In the proposed feature extraction scheme, the principal component analysis (PCA) is applied to the dyadic wavelet transform (DWT) of the ECG signal to extract morphological ECG features, which are then combined with the temporal features to form a resultant efficient feature vector. For the classification scheme, we selected the feed-forward artificial neural networks (ANNs) optimally designed by the multi-dimensional particle swarm optimization (MD-PSO) technique, which evolves the structure and weights of the network specifically for each patient. Training data for the ANN classifier include both global (total of 150 representative beats randomly sampled from each class in selected training files) and local (the first 5 min of a patient's ECG recording) training patterns. Simulation results using 40 files in the MIT/BIH arrhythmia database achieved high average accuracy of 97% for differentiating normal, PVC, and other beats.
Citation - Scopus: 4
Biosignal Time-Series Analysis
(Elsevier, 2022) Kiranyaz S.; İnce, Türker; Chowdhury M.E.H.; Degerli A.; Gabbouj, Moncef; Kiranyaz, Serkan; Chowdhury, Muhammad E. H.; Degerli, Aysen
In this chapter, recent state-of-the-art techniques in biosignal time-series analysis will be presented. We shall start with the problem of patient-specific ECG beat classification where the objective is to discriminate the arrhythmic beats from the normal (healthy) beats of an individual patient. So, we will answer the ultimate question of how to design person-specific, real-time, and accurate monitoring of ECG signals. We shall then move on to the recent solution of a related problem, an early warning system that can alert an individual the instant his/her heart deviates from its normal rhythm. This is a far challenging problem since the detection of the arrhythmia beats should be performed without knowing them. © 2022 Elsevier Inc. All rights reserved.
Citation - WoS: 29
Citation - Scopus: 37
Blind Ecg Restoration by Operational Cycle-Gans
(IEEE-Inst Electrical Electronics Engineers Inc, 2022) Kiranyaz, Serkan; Devecioglu, Ozer Can; İnce, Türker; Malik, Junaid; Chowdhury, Muhammad; Hamid, Tahir; Mazhar, Rashid; Gabbouj, Moncef
Objective: ECG recordings often suffer from a set of artifacts with varying types, severities, and durations, and this makes an accurate diagnosis by machines or medical doctors difficult and unreliable. Numerous studies have proposed ECG denoising; however, they naturally fail to restore the actual ECG signal corrupted with such artifacts due to their simple and naive noise model. In this pilot study, we propose a novel approach for blind ECG restoration using cycle-consistent generative adversarial networks (Cycle-GANs) where the quality of the signal can be improved to a clinical level ECG regardless of the type and severity of the artifacts corrupting the signal. Methods: To further boost the restoration performance, we propose 1D operational Cycle-GANs with the generative neuron model. Results: The proposed approach has been evaluated extensively using one of the largest benchmark ECG datasets from the China Physiological Signal Challenge (CPSC-2020) with more than one million beats. Besides the quantitative and qualitative evaluations, a group of cardiologists performed medical evaluations to validate the quality and usability of the restored ECG, especially for an accurate arrhythmia diagnosis. Significance: As a pioneer study in ECG restoration, the corrupted ECG signals can be restored to clinical level quality. Conclusion: By means of the proposed ECG restoration, the ECG diagnosis accuracy and performance can significantly improve.
Citation - WoS: 33
Citation - Scopus: 37
Classification and Retrieval on Macroinvertebrate Image Databases
(Pergamon-Elsevier Science Ltd, 2011) Kiranyaz, Serkan; İnce, Türker; Pulkkinen, Jenni; Gabbouj, Moncef; Arje, Johanna; Karkkainen, Salme; Tirronen, Ville
Aquatic ecosystems are continuously threatened by a growing number of human induced changes. Macroinvertebrate biomonitoring is particularly efficient in pinpointing the cause-effect structure between slow and subtle changes and their detrimental consequences in aquatic ecosystems. The greatest obstacle to implementing efficient biomonitoring is currently the cost-intensive human expert taxonomic identification of samples. While there is evidence that automated recognition techniques can match human taxa identification accuracy at greatly reduced costs, so far the development of automated identification techniques for aquatic organisms has been minimal. In this paper, we focus on advancing classification and data retrieval that are instrumental when processing large macroinvertebrate image datasets. To accomplish this for routine biomonitoring, in this paper we shall investigate the feasibility of automated river macroinvertebrate classification and retrieval with high precision. Besides the state-of-the-art classifiers such as Support Vector Machines (SVMs) and Bayesian Classifiers (BCs), the focus is particularly drawn on feed-forward artificial neural networks (ANNs), namely multilayer perceptrons (MLPs) and radial basis function networks (RBFNs). Since both ANN types have been proclaimed superior by different investigations even for the same benchmark problems, we shall first show that the main reason for this ambiguity lies in the static and rather poor comparison methodologies applied in most earlier works. Especially the most common drawback occurs due to the limited evaluation of the ANN performances over just one or few network architecture(s). Therefore, in this study, an extensive evaluation of each classifier performance over an ANN architecture space is performed. The best classifier among all, which is trained over a dataset of river macroinvertebrate specimens, is then used in the MUVIS framework for the efficient search and retrieval of particular macroinvertebrate peculiars. Classification and retrieval results present high accuracy and can match an experts' ability for taxonomic identification. (C) 2011 Elsevier Ltd. All rights reserved.
Citation - WoS: 22
Citation - Scopus: 26
Classification of Polarimetric Sar Images Using Compact Convolutional Neural Networks
(Taylor & Francis Ltd, 2021) Ahishali, Mete; Kiranyaz, Serkan; İnce, Türker; Gabbouj, Moncef
Classification of polarimetric synthetic aperture radar (PolSAR) images is an active research area with a major role in environmental applications. The traditional Machine Learning (ML) methods proposed in this domain generally focus on utilizing highly discriminative features to improve the classification performance, but this task is complicated by the well-known curse of dimensionality phenomena. Other approaches based on deep Convolutional Neural Networks (CNNs) have certain limitations and drawbacks, such as high computational complexity, an unfeasibly large training set with ground-truth labels, and special hardware requirements. In this work, to address the limitations of traditional ML and deep CNN-based methods, a novel and systematic classification framework is proposed for the classification of PolSAR images, based on a compact and adaptive implementation of CNNs using a sliding-window classification approach. The proposed approach has three advantages. First, there is no requirement for an extensive feature extraction process. Second, it is computationally efficient due to utilized compact configurations. In particular, the proposed compact and adaptive CNN model is designed to achieve the maximum classification accuracy with minimum training and computational complexity. This is of considerable importance considering the high costs involved in labeling in PolSAR classification. Finally, the proposed approach can perform classification using smaller window sizes than deep CNNs. Experimental evaluations have been performed over the most commonly used four benchmark PolSAR images: AIRSAR L-Band and RADARSAT-2 C-Band data of San Francisco Bay and Flevoland areas. Accordingly, the best obtained overall accuracies range between 92.33-99.39% for these benchmark study sites.
Citation - Scopus: 4
Classification of Polarimetric Sar Images Using Evolutionary Rbf Networks
(2010) İnce, Türker; Kiranyaz S.; Gabbouj, Moncef
This paper proposes an evolutionary RBF network classifier for polarimetric synthetic aperture radar ( SAR) images. The proposed feature extraction process utilizes the full covariance matrix, the gray level co-occurrence matrix (GLCM) based texture features, and the backscattering power (Span) combined with the H/?/A decomposition, which are projected onto a lower dimensional feature space using principal component analysis. An experimental study is performed using the fully polarimetric San Francisco Bay data set acquired by the NASA/Jet Propulsion Laboratory Airborne SAR (AIRSAR) at L-band to evaluate the performance of the proposed classifier. Classification results (in terms of confusion matrix, overall accuracy and classification map) compared to the Wishart and a recent NN-based classifiers demonstrate the effectiveness of the proposed algorithm. © 2010 IEEE.
Citation - WoS: 15
Citation - Scopus: 19
Collective Network of Binary Classifier Framework for Polarimetric Sar Image Classification: an Evolutionary Approach
(IEEE-Inst Electrical Electronics Engineers Inc, 2012) Kiranyaz, Serkan; İnce, Türker; Uhlmann, Stefan; Gabbouj, Moncef
Terrain classification over polarimetric synthetic aperture radar (SAR) images has been an active research field where several features and classifiers have been proposed up to date. However, some key questions, e.g., 1) how to select certain features so as to achieve highest discrimination over certain classes?, 2) how to combine them in the most effective way?, 3) which distance metric to apply?, 4) how to find the optimal classifier configuration for the classification problem in hand?, 5) how to scale/adapt the classifier if large number of classes/features are present?, and finally, 6) how to train the classifier efficiently to maximize the classification accuracy?, still remain unanswered. In this paper, we propose a collective network of (evolutionary) binary classifier (CNBC) framework to address all these problems and to achieve high classification performance. The CNBC framework adapts a Divide and Conquer type approach by allocating several NBCs to discriminate each class and performs evolutionary search to find the optimal BC in each NBC. In such an (incremental) evolution session, the CNBC body can further dynamically adapt itself with each new incoming class/feature set without a full-scale retraining or reconfiguration. Both visual and numerical performance evaluations of the proposed framework over two benchmark SAR images demonstrate its superiority and a significant performance gap against several major classifiers in this field.
Citation - Scopus: 5
Collective Network of Evolutionary Binary Classifiers for Content-Based Image Retrieval
(2011) Kiranyaz S.; Uhlmann S.; Pulkkinen J.; Gabbouj, Moncef; İnce, Türker
The content-based image retrieval (CBIR) has been an active research field for which several feature extraction, classification and retrieval techniques have been proposed up to date. However, when the database size grows larger, it is a common fact that the overall retrieval performance significantly deteriorates. In this paper, we propose collective network of (evolutionary) binary classifiers (CNBC) framework to achieve a high retrieval performance even though the training (ground truth) data may not be entirely present from the beginning and thus the system can only be trained incrementally. The CNBC framework basically adopts a "Divide and Conquer" type approach by allocating several networks of binary classifiers (NBCs) to discriminate each class and performs evolutionary search to find the optimal binary classifier (BC) in each NBC. In such an evolution session, the CNBC body can further dynamically adapt itself with each new incoming class/feature set without a full-scale re-training or re-configuration. Both visual and numerical performance evaluations of the proposed framework over benchmark image databases demonstrate its scalability; and a significant performance improvement is achieved over traditional retrieval techniques. © 2011 IEEE.
Citation - Scopus: 291
Convolutional Neural Networks for Patient-Specific Ecg Classification
(Institute of Electrical and Electronics Engineers Inc., 2015) Kiranyaz S.; İnce, Türker; Hamila R.; Gabbouj, Moncef
We propose a fast and accurate patient-specific electrocardiogram (ECG) classification and monitoring system using an adaptive implementation of 1D Convolutional Neural Networks (CNNs) that can fuse feature extraction and classification into a unified learner. In this way, a dedicated CNN will be trained for each patient by using relatively small common and patient-specific training data and thus it can also be used to classify long ECG records such as Holter registers in a fast and accurate manner. Alternatively, such a solution can conveniently be used for real-time ECG monitoring and early alert system on a light-weight wearable device. The experimental results demonstrate that the proposed system achieves a superior classification performance for the detection of ventricular ectopic beats (VEB) and supraventricular ectopic beats (SVEB). © 2015 IEEE.
Citation - WoS: 19
Citation - Scopus: 24
Dual and Single Polarized Sar Image Classification Using Compact Convolutional Neural Networks
(Mdpi, 2019) Ahishali, Mete; Kiranyaz, Serkan; İnce, Türker; Gabbouj, Moncef
Accurate land use/land cover classification of synthetic aperture radar (SAR) images plays an important role in environmental, economic, and nature related research areas and applications. When fully polarimetric SAR data is not available, single- or dual-polarization SAR data can also be used whilst posing certain difficulties. For instance, traditional Machine Learning (ML) methods generally focus on finding more discriminative features to overcome the lack of information due to single- or dual-polarimetry. Beside conventional ML approaches, studies proposing deep convolutional neural networks (CNNs) come with limitations and drawbacks such as requirements of massive amounts of data for training and special hardware for implementing complex deep networks. In this study, we propose a systematic approach based on sliding-window classification with compact and adaptive CNNs that can overcome such drawbacks whilst achieving state-of-the-art performance levels for land use/land cover classification. The proposed approach voids the need for feature extraction and selection processes entirely, and perform classification directly over SAR intensity data. Furthermore, unlike deep CNNs, the proposed approach requires neither a dedicated hardware nor a large amount of data with ground-truth labels. The proposed systematic approach is designed to achieve maximum classification accuracy on single and dual-polarized intensity data with minimum human interaction. Moreover, due to its compact configuration, the proposed approach can process such small patches which is not possible with deep learning solutions. This ability significantly improves the details in segmentation masks. An extensive set of experiments over two benchmark SAR datasets confirms the superior classification performance and efficient computational complexity of the proposed approach compared to the competing methods.
Citation - Scopus: 1
Dynamic Data Clustering Using Stochastic Approximation Driven Multi-Dimensional Particle Swarm Optimization
(Springer Verlag, 2010) Kiranyaz S.; İnce, Türker; Gabbouj, Moncef
With an ever-growing attention Particle Swarm Optimization (PSO) has found many application areas for many challenging optimization problems. It is, however, a known fact that PSO has a severe drawback in the update of its global best (gbest) particle, which has a crucial role of guiding the rest of the swarm. In this paper, we propose two efficient solutions to remedy this problem using a stochastic approximation (SA) technique. For this purpose we use simultaneous perturbation stochastic approximation (SPSA), which is applied only to the gbest (not to the entire swarm) for a low-cost solution. Since the problem of poor gbest update persists in the recently proposed extension of PSO, called multi-dimensional PSO (MD-PSO), two distinct SA approaches are then integrated into MD-PSO and tested over a set of unsupervised data clustering applications. Experimental results show that the proposed approaches significantly improved the quality of the MD-PSO clustering as measured by a validity index function. Furthermore, the proposed approaches are generic as they can be used with other PSO variants and applicable to a wide range of problems. © 2010 Springer-Verlag Berlin Heidelberg.
Citation - WoS: 24
Citation - Scopus: 33
Early Bearing Fault Diagnosis of Rotating Machinery by 1d Self-Organized Operational Neural Networks
(IEEE-Inst Electrical Electronics Engineers Inc, 2021) İnce, Türker; Malik, Junaid; Devecioglu, Ozer Can; Kiranyaz, Serkan; Avcı, Onur; Eren, Levent; Gabbouj, Moncef
Preventive maintenance of modern electric rotating machinery (RM) is critical for ensuring reliable operation, preventing unpredicted breakdowns and avoiding costly repairs. Recently many studies investigated machine learning monitoring methods especially based on Deep Learning networks focusing mostly on detecting bearing faults; however, none of them addressed bearing fault severity classification for early fault diagnosis with high enough accuracy. 1D Convolutional Neural Networks (CNNs) have indeed achieved good performance for detecting RM bearing faults from raw vibration and current signals but did not classify fault severity. Furthermore, recent studies have demonstrated the limitation in terms of learning capability of conventional CNNs attributed to the basic underlying linear neuron model. Recently, Operational Neural Networks (ONNs) were proposed to enhance the learning capability of CNN by introducing non-linear neuron models and further heterogeneity in the network configuration. In this study, we propose 1D Self-organized ONNs (Self-ONNs) with generative neurons for bearing fault severity classification and providing continuous condition monitoring. Experimental results over the benchmark NSF/IMS bearing vibration dataset using both x- and y-axis vibration signals for inner race and rolling element faults demonstrate that the proposed 1D Self-ONNs achieve significant performance gap against the state-of-the-art (1D CNNs) with similar computational complexity.
Citation - WoS: 5
Citation - Scopus: 5
The Effect of Automated Taxa Identification Errors on Biological Indices
(Pergamon-Elsevier Science Ltd, 2017) Arje, Johanna; Karkkainen, Salme; Meissner, Kristian; Iosifidis, Alexandros; İnce, Türker; Gabbouj, Moncef; Kiranyaz, Serkan
In benthic macroinvertebrate biomonitoring systems, the target is to determine the status of ecosystems based on several biological indices. To increase cost-efficiency, computer-based taxa identification for image data has recently been developed. Taxa identification errors can, however, have strong effects on the indices and thus on the determination of the ecological status. In order to shift the biomonitoring process towards automated expert systems, we need a clear understanding on the bias caused by automation. In this paper, we examine eleven classification methods in the case of macroinvertebrate image data and show how their classification errors propagate into different biological indices. We evaluate 14 richness, diversity, dominance and similarity indices commonly used in biomonitoring. Besides the error rate of the classification method, we discuss the potential effect of different types of identification errors. Finally, we provide recommendations on indices that are least affected by the automatic identification errors and could be used in automated biomonitoring. (C) 2016 Elsevier Ltd. All rights reserved.
Citation - WoS: 30
Citation - Scopus: 39
Evaluation of Global and Local Training Techniques Over Feed-Forward Neural Network Architecture Spaces for Computer-Aided Medical Diagnosis
(Pergamon-Elsevier Science Ltd, 2010) İnce, Türker; Kiranyaz, Serkan; Pulkkinen, Jenni; Gabbouj, Moncef
In this paper, we investigate the performance of global vs. local techniques applied to the training of neural network classifiers for solving medical diagnosis problems. The presented methodology of the investigation involves systematic and exhaustive evaluation of the classifier performance over a neural network architecture space and with respect to training depth for a particular problem. In this study, the architecture space is defined over feed-forward, fully-connected artificial neural networks (ANNs) which have been widely used in computer-aided decision support systems in medical domain, and for which two popular neural network training methods are explored: conventional backpropagation (BP) and particle swarm optimization (PSO). Both training techniques are compared in terms of classification performance over three medical diagnosis problems (breast cancer, heart disease, and diabetes) from Pro-ben1 benchmark dataset and computational and architectural analysis are performed for an extensive assessment. The results clearly demonstrate that it is not possible to compare and evaluate the performance of the two algorithms over a single network and with a fixed set of training parameters, as most of the earlier work in this field has been carried out, since training and test classification performances vary significantly and depend directly on the network architecture, the training depth and method used and the available dataset. We, therefore, show that an extensive evaluation method such as the one proposed in this paper is basically needed to obtain a reliable and detailed performance assessment, in that, we can conclude that the PSO algorithm has usually a better generalization ability across the architecture space whereas BP can occasionally provide better training and/or test classification performance for some network configurations. Furthermore, we can in general say that the PSO, as a global training algorithm, is capable of achieving minimum test classification errors regardless of the training depth, i.e. shallow or deep, and its average classification performance shows less variations with respect to network architecture. In terms of computational complexity, BP is in general superior to PSO for the entire architecture space used. (C) 2010 Elsevier Ltd. All rights reserved.
Citation - WoS: 190
Citation - Scopus: 246
Evolutionary Artificial Neural Networks by Multi-Dimensional Particle Swarm Optimization
(Pergamon-Elsevier Science Ltd, 2009) Kiranyaz, Serkan; İnce, Türker; Yildirim, Alper; Gabbouj, Moncef
In this paper, we propose a novel technique for the automatic design of Artificial Neural Networks (ANNs) by evolving to the optimal network configuration(s) within an architecture space. It is entirely based on a multi-dimensional Particle Swarm Optimization (MD PSO) technique, which re-forms the native structure of swarm particles in such a way that they can make inter-dimensional passes with a dedicated dimensional PSO process. Therefore, in a multidimensional search space where the optimum dimension is unknown, swarm particles can seek both positional and dimensional optima. This eventually removes the necessity of setting a fixed dimension a priori, which is a common drawback for the family of swarm optimizers. With the proper encoding of the network configurations and parameters into particles, MID PSO can then seek the positional optimum in the error space and the dimensional optimum in the architecture space. The optimum dimension converged at the end of a MD PSO process corresponds to a unique ANN configuration where the network parameters (connections, weights and biases) can then be resolved from the positional optimum reached on that dimension. In addition to this, the proposed technique generates a ranked list of network configurations, from the best to the worst. This is indeed a crucial piece of information, indicating what potential configurations can be alternatives to the best one, and which configurations should not be used at all for a particular problem. In this study, the architecture space is defined over feed-forward, fully-connected ANNs so as to use the conventional techniques such as back-propagation and some other evolutionary methods in this field. The proposed technique is applied over the most challenging synthetic problems to test its optimality on evolving networks and over the benchmark problems to test its generalization capability as well as to make comparative evaluations with the several competing techniques. The experimental results show that the MD PSO evolves to optimum or near-optimum networks in general and has a superior generalization capability. Furthermore, the MID PSO naturally favors a low-dimension solution when it exhibits a competitive performance with a high dimension counterpart and such a native tendency eventually yields the evolution process to the compact network configurations in the architecture space rather than the complex ones, as long as the optimality prevails. (C) 2009 Elsevier Ltd. All rights reserved.
Citation - WoS: 25
Citation - Scopus: 29
Evolutionary Rbf Classifier for Polarimetric Sar Images
(Pergamon-Elsevier Science Ltd, 2012) İnce, Türker; Kiranyaz, Serkan; Gabbouj, Moncef
In this paper, a robust radial basis function (RBF) network based classifier is proposed for polarimetric synthetic aperture radar (SAR) images. The proposed feature extraction process utilizes the covariance matrix elements, the H/alpha/A decomposition based features combined with the backscattering power (span), and the gray level co-occurrence matrix (GLCM) based texture features, which are projected onto a lower dimensional feature space using principal components analysis. For the classifier training, both conventional backpropagation (BP) and multidimensional particle swarm optimization (MD-PSO) based dynamic clustering are explored. By combining complete polarimetric covariance matrix and eigenvalue decomposition based pixel values with textural information (contrast, correlation, energy, and homogeneity) in the feature set, and employing automated evolutionary RBF classifier for the pattern recognition unit, the overall classification performance is shown to be significantly improved. An experimental study is performed using the fully polarimetric San Francisco Bay and Flevoland data sets acquired by the NASA/Jet Propulsion Laboratory Airborne SAR (AIRSAR) at L-band to evaluate the performance of the proposed classifier. Classification results (in terms of confusion matrix, overall accuracy and classification map) compared with the major state of the art algorithms demonstrate the effectiveness of the proposed RBF network classifier. (C) 2011 Elsevier Ltd. All rights reserved.