1
|
Kaur M, Singh D, Kumar V, Lee HN. MLNet: Metaheuristics-Based Lightweight Deep Learning Network for Cervical Cancer Diagnosis. IEEE J Biomed Health Inform 2023; 27:5004-5014. [PMID: 36399582 DOI: 10.1109/jbhi.2022.3223127] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/06/2023]
Abstract
One of the leading causes of cancer-related deaths among women is cervical cancer. Early diagnosis and treatment can minimize the complications of this cancer. Recently, researchers have designed and implemented many deep learning-based automated cervical cancer diagnosis models. However, the majority of these models suffer from over-fitting, parameter tuning, and gradient vanishing problems. To overcome these problems, in this paper a metaheuristics-based lightweight deep learning network (MLNet) is proposed. Initially, the hyper-parameters tuning problem of convolutional neural network (CNN) is defined as a multi-objective problem. Particle swarm optimization (PSO) is used to optimally define the CNN architecture. Thereafter, Dynamically hybrid niching differential evolution (DHDE) is utilized to optimize the hyper-parameters of CNN layers. Each particle of PSO and solution of DHDE together represent the possible CNN configuration. F-score is used as a fitness function. The proposed MLNet is trained and validated on three benchmark cervical cancer datasets. On the Herlev dataset, MLNet outperforms the existing models in terms of accuracy, f-measure, sensitivity, specificity, and precision by 1.6254%, 1.5178%, 1.5780%, 1.7145%, and 1.4890%, respectively. Also, on the SIPaKMeD dataset, MLNet achieves better performance than the existing models in terms of accuracy, f-measure, sensitivity, specificity, and precision by 2.1250%, 2.2455%, 1.9074%, 1.9258%, and 1.8975%, respectively. Finally, on the Mendeley LBC dataset, MLNet achieves better performance than the competitive models in terms of accuracy, f-measure, sensitivity, specificity, and precision by 1.4680%, 1.5845%, 1.3582%, 1.3926%, and 1.4125%, respectively.
Collapse
|
2
|
Su X, Sun Y, Liu H, Lang Q, Zhang Y, Zhang J, Wang C, Chen Y. An innovative ensemble model based on deep learning for predicting COVID-19 infection. Sci Rep 2023; 13:12322. [PMID: 37516796 PMCID: PMC10387055 DOI: 10.1038/s41598-023-39408-8] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/26/2023] [Accepted: 07/25/2023] [Indexed: 07/31/2023] Open
Abstract
Nowadays, global public health crises are occurring more frequently, and accurate prediction of these diseases can reduce the burden on the healthcare system. Taking COVID-19 as an example, accurate prediction of infection can assist experts in effectively allocating medical resources and diagnosing diseases. Currently, scholars worldwide use single model approaches or epidemiology models more often to predict the outbreak trend of COVID-19, resulting in poor prediction accuracy. Although a few studies have employed ensemble models, there is still room for improvement in their performance. In addition, there are only a few models that use the laboratory results of patients to predict COVID-19 infection. To address these issues, research efforts should focus on improving disease prediction performance and expanding the use of medical disease prediction models. In this paper, we propose an innovative deep learning model Whale Optimization Convolutional Neural Networks (CNN), Long-Short Term Memory (LSTM) and Artificial Neural Network (ANN) called WOCLSA which incorporates three models ANN, CNN and LSTM. The WOCLSA model utilizes the Whale Optimization Algorithm to optimize the neuron number, dropout and batch size parameters in the integrated model of ANN, CNN and LSTM, thereby finding the global optimal solution parameters. WOCLSA employs 18 patient indicators as predictors, and compares its results with three other ensemble deep learning models. All models were validated with train-test split approaches. We evaluate and compare our proposed model and other models using accuracy, F1 score, recall, AUC and precision metrics. Through many studies and tests, our results show that our prediction models can identify patients with COVID-19 infection at the AUC of 91%, 91%, and 93% respectively. Other prediction results achieve a respectable accuracy of 92.82%, 92.79%, and 91.66% respectively, f1-score of 93.41%, 92.79%, and 92.33% respectively, precision of 93.41%, 92.79%, and 92.33% respectively, recall of 93.41%, 92.79%, and 92.33% respectively. All of these exceed 91%, surpassing those of comparable models. The execution time of WOCLSA is also an advantage. Therefore, the WOCLSA ensemble model can be used to assist in verifying laboratory research results and predict and to judge various diseases in public health events.
Collapse
Affiliation(s)
- Xiaoying Su
- School of Jilin Emergency Management, Changchun Institute of Technology, Changchun, 130021, China
| | - Yanfeng Sun
- College of Computer Science and Technology, Jilin University, Changchun, 130012, China
| | - Hongxi Liu
- School of Jilin Emergency Management, Changchun Institute of Technology, Changchun, 130021, China
| | - Qiuling Lang
- School of Jilin Emergency Management, Changchun Institute of Technology, Changchun, 130021, China
| | - Yichen Zhang
- School of Jilin Emergency Management, Changchun Institute of Technology, Changchun, 130021, China
| | - Jiquan Zhang
- School of Environment, Northeast Normal University, Changchun, 130024, China
| | - Chaoyong Wang
- School of Jilin Emergency Management, Changchun Institute of Technology, Changchun, 130021, China.
| | - Yanan Chen
- School of Jilin Emergency Management, Changchun Institute of Technology, Changchun, 130021, China
| |
Collapse
|
3
|
He X, Shan W, Zhang R, Heidari AA, Chen H, Zhang Y. Improved Colony Predation Algorithm Optimized Convolutional Neural Networks for Electrocardiogram Signal Classification. Biomimetics (Basel) 2023; 8:268. [PMID: 37504156 PMCID: PMC10377160 DOI: 10.3390/biomimetics8030268] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/13/2023] [Revised: 06/18/2023] [Accepted: 06/18/2023] [Indexed: 07/29/2023] Open
Abstract
Recently, swarm intelligence algorithms have received much attention because of their flexibility for solving complex problems in the real world. Recently, a new algorithm called the colony predation algorithm (CPA) has been proposed, taking inspiration from the predatory habits of groups in nature. However, CPA suffers from poor exploratory ability and cannot always escape solutions known as local optima. Therefore, to improve the global search capability of CPA, an improved variant (OLCPA) incorporating an orthogonal learning strategy is proposed in this paper. Then, considering the fact that the swarm intelligence algorithm can go beyond the local optimum and find the global optimum solution, a novel OLCPA-CNN model is proposed, which uses the OLCPA algorithm to tune the parameters of the convolutional neural network. To verify the performance of OLCPA, comparison experiments are designed to compare with other traditional metaheuristics and advanced algorithms on IEEE CEC 2017 benchmark functions. The experimental results show that OLCPA ranks first in performance compared to the other algorithms. Additionally, the OLCPA-CNN model achieves high accuracy rates of 97.7% and 97.8% in classifying the MIT-BIH Arrhythmia and European ST-T datasets.
Collapse
Affiliation(s)
- Xinxin He
- School of Emergency Management, Institute of Disaster Prevention, Sanhe 065201, China
| | - Weifeng Shan
- School of Emergency Management, Institute of Disaster Prevention, Sanhe 065201, China
| | - Ruilei Zhang
- School of Emergency Management, Institute of Disaster Prevention, Sanhe 065201, China
| | - Ali Asghar Heidari
- School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran, Tehran 1417935840, Iran
| | - Huiling Chen
- Institute of Big Data and Information Technology, Wenzhou University, Wenzhou 325000, China
| | - Yudong Zhang
- School of Computing and Mathematical Sciences, University of Leicester, Leicester LE1 7RH, UK
| |
Collapse
|
4
|
Huang J, Zeng G, Geng G, Weng J, Lu K. SOPA‐GA‐CNN: Synchronous optimisation of parameters and architectures by genetic algorithms with convolutional neural network blocks for securing Industrial Internet‐of‐Things. IET CYBER-SYSTEMS AND ROBOTICS 2023. [DOI: 10.1049/csy2.12085] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 03/29/2023] Open
|
5
|
Appadurai JP, G S, Prabhu Kavin B, C K, Lai WC. Multi-Process Remora Enhanced Hyperparameters of Convolutional Neural Network for Lung Cancer Prediction. Biomedicines 2023; 11:biomedicines11030679. [PMID: 36979657 PMCID: PMC10045623 DOI: 10.3390/biomedicines11030679] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/28/2022] [Revised: 01/30/2023] [Accepted: 02/08/2023] [Indexed: 03/30/2023] Open
Abstract
In recent years, lung cancer prediction is an essential topic for reducing the death rate of humans. In the literature section, some papers are reviewed that reduce the accuracy level during the prediction stage. Hence, in this paper, we develop a Multi-Process Remora Optimized Hyperparameters of Convolutional Neural Network (MPROH-CNN) aimed at lung cancer prediction. The proposed technique can be utilized to detect the CT images of the human lung. The proposed technique proceeds with four phases, including pre-processing, feature extraction and classification. Initially, the databases are collected from the open-source system. After that, the collected CT images contain unwanted noise, which affects classification efficiency. So, the pre-processing techniques can be considered to remove unwanted noise from the input images, such as filtering and contrast enhancement. Following that, the essential features are extracted with the assistance of feature extraction techniques such as histogram, texture and wavelet. The extracted features are utilized to classification stage. The proposed classifier is a combination of the Remora Optimization Algorithm (ROA) and Convolutional Neural Network (CNN). In the CNN, the ROA is utilized for multi process optimization such as structure optimization and hyperparameter optimization. The proposed methodology is implemented in MATLAB and performances are evaluated by utilized performance matrices such as accuracy, precision, recall, specificity, sensitivity and F_Measure. To validate the projected approach, it is compared with the traditional techniques CNN, CNN-Particle Swarm Optimization (PSO) and CNN-Firefly Algorithm (FA), respectively. From the analysis, the proposed method achieved a 0.98 accuracy level in the lung cancer prediction.
Collapse
Affiliation(s)
- Jothi Prabha Appadurai
- Computer Science and Engineering Department, Kakatiya Institute of Technology and Science, Warangal 506015, Telangana, India
| | - Suganeshwari G
- School of Computer Science and Engineering, Vellore Institute of Technology, Chennai 600127, Tamil Nadu, India
| | - Balasubramanian Prabhu Kavin
- Department of Data Science and Business Systems, College of Engineering and Technology, SRM Institute of Science and Technology, SRM Nagar, Chengalpattu District, Chennai 603203, Tamil Nadu, India
| | - Kavitha C
- Department of Computer Science and Engineering, Sathyabama Institute of Science and Technology, Chennai 600119, Tamil Nadu, India
| | - Wen-Cheng Lai
- Bachelor Program in Industrial Projects, National Yunlin University of Science and Technology, Douliu 640301, Taiwan
- Department Electronic Engineering, National Yunlin University of Science and Technology, Douliu 640301, Taiwan
| |
Collapse
|
6
|
Sathya T, Sudha S. OQCNN: optimal quantum convolutional neural network for classification of facial expression. Neural Comput Appl 2023. [DOI: 10.1007/s00521-022-08161-w] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 02/07/2023]
|
7
|
Xie J, Chen R, Bhattacharyya SS. A parameter-optimization framework for neural decoding systems. Front Neuroinform 2023; 17:938689. [PMID: 36817969 PMCID: PMC9932190 DOI: 10.3389/fninf.2023.938689] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/07/2022] [Accepted: 01/06/2023] [Indexed: 02/05/2023] Open
Abstract
Real-time neuron detection and neural activity extraction are critical components of real-time neural decoding. They are modeled effectively in dataflow graphs. However, these graphs and the components within them in general have many parameters, including hyper-parameters associated with machine learning sub-systems. The dataflow graph parameters induce a complex design space, where alternative configurations (design points) provide different trade-offs involving key operational metrics including accuracy and time-efficiency. In this paper, we propose a novel optimization framework that automatically configures the parameters in different neural decoders. The proposed optimization framework is evaluated in depth through two case studies. Significant performance improvement in terms of accuracy and efficiency is observed in both case studies compared to the manual parameter optimization that was associated with the published results of those case studies. Additionally, we investigate the application of efficient multi-threading strategies to speed-up the running time of our parameter optimization framework. Our proposed optimization framework enables efficient and effective estimation of parameters, which leads to more powerful neural decoding capabilities and allows researchers to experiment more easily with alternative decoding models.
Collapse
Affiliation(s)
- Jing Xie
- Department of Electrical and Computer Engineering, University of Maryland at College Park, College Park, MD, United States
| | - Rong Chen
- Department of Diagnostic Radiology and Nuclear Medicine, University of Maryland at Baltimore, Baltimore, MD, United States
| | - Shuvra S. Bhattacharyya
- Department of Electrical and Computer Engineering, University of Maryland at College Park, College Park, MD, United States
- Institute for Advanced Computer Studies (UMIACS), University of Maryland at College Park, College Park, MD, United States
| |
Collapse
|
8
|
Shami TM, Mirjalili S, Al-Eryani Y, Daoudi K, Izadi S, Abualigah L. Velocity pausing particle swarm optimization: a novel variant for global optimization. Neural Comput Appl 2023. [DOI: 10.1007/s00521-022-08179-0] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/08/2023]
Abstract
AbstractParticle swarm optimization (PSO) is one of the most well-regard metaheuristics with remarkable performance when solving diverse optimization problems. However, PSO faces two main problems that degrade its performance: slow convergence and local optima entrapment. In addition, the performance of this algorithm substantially degrades on high-dimensional problems. In the classical PSO, particles can move in each iteration with either slower or faster speed. This work proposes a novel idea called velocity pausing where particles in the proposed velocity pausing PSO (VPPSO) variant are supported by a third movement option that allows them to move with the same velocity as they did in the previous iteration. As a result, VPPSO has a higher potential to balance exploration and exploitation. To avoid the PSO premature convergence, VPPSO modifies the first term of the PSO velocity equation. In addition, the population of VPPSO is divided into two swarms to maintain diversity. The performance of VPPSO is validated on forty three benchmark functions and four real-world engineering problems. According to the Wilcoxon rank-sum and Friedman tests, VPPSO can significantly outperform seven prominent algorithms on most of the tested functions on both low- and high-dimensional cases. Due to its superior performance in solving complex high-dimensional problems, VPPSO can be applied to solve diverse real-world optimization problems. Moreover, the velocity pausing concept can be easily integrated with new or existing metaheuristic algorithms to enhance their performances. The Matlab code of VPPSO is available at: https://uk.mathworks.com/matlabcentral/fileexchange/119633-vppso.
Collapse
|
9
|
Saad MH, Hashima S, Sayed W, El-Shazly EH, Madian AH, Fouda MM. Early Diagnosis of COVID-19 Images Using Optimal CNN Hyperparameters. Diagnostics (Basel) 2022; 13:diagnostics13010076. [PMID: 36611368 PMCID: PMC9818649 DOI: 10.3390/diagnostics13010076] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/24/2022] [Revised: 12/13/2022] [Accepted: 12/19/2022] [Indexed: 12/29/2022] Open
Abstract
Coronavirus disease (COVID-19) is a worldwide epidemic that poses substantial health hazards. However, COVID-19 diagnostic test sensitivity is still restricted due to abnormalities in specimen processing. Meanwhile, optimizing the highly defined number of convolutional neural network (CNN) hyperparameters (hundreds to thousands) is a useful direction to improve its overall performance and overcome its cons. Hence, this paper proposes an optimization strategy for obtaining the optimal learning rate and momentum of a CNN's hyperparameters using the grid search method to improve the network performance. Therefore, three alternative CNN architectures (GoogleNet, VGG16, and ResNet) were used to optimize hyperparameters utilizing two different COVID-19 radiography data sets (Kaggle (X-ray) and China national center for bio-information (CT)). These architectures were tested with/without optimizing the hyperparameters. The results confirm effective disease classification using the CNN structures with optimized hyperparameters. Experimental findings indicate that the new technique outperformed the previous in terms of accuracy, sensitivity, specificity, recall, F-score, false positive and negative rates, and error rate. At epoch 25, the optimized Resnet obtained high classification accuracy, reaching 98.98% for X-ray images and 98.78% for CT images.
Collapse
Affiliation(s)
- Mohamed H. Saad
- Radiation Engineering Department, National Center for Radiation Research and Technology (NCRRT), Egyptian Atomic Energy Authority, Cairo 11787, Egypt
| | - Sherief Hashima
- Engineering Department, Nuclear Research Center (NRC), Egyptian Atomic Energy Authority, Cairo 13759, Egypt
- Correspondence: ; Tel.: +20-10-94230077
| | - Wessam Sayed
- Radiation Engineering Department, National Center for Radiation Research and Technology (NCRRT), Egyptian Atomic Energy Authority, Cairo 11787, Egypt
| | - Ehab H. El-Shazly
- Radiation Engineering Department, National Center for Radiation Research and Technology (NCRRT), Egyptian Atomic Energy Authority, Cairo 11787, Egypt
| | - Ahmed H. Madian
- Radiation Engineering Department, National Center for Radiation Research and Technology (NCRRT), Egyptian Atomic Energy Authority, Cairo 11787, Egypt
| | - Mostafa M. Fouda
- Department of Electrical and Computer Engineering, Idaho State University, Pocatello, ID 83209, USA
- Department of Electrical Engineering, Faculty of Engineering at Shoubra, Benha University, Cairo 11672, Egypt
| |
Collapse
|
10
|
Han X, Hu Z, Wang S, Zhang Y. A Survey on Deep Learning in COVID-19 Diagnosis. J Imaging 2022; 9:jimaging9010001. [PMID: 36662099 PMCID: PMC9866755 DOI: 10.3390/jimaging9010001] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/15/2022] [Revised: 12/05/2022] [Accepted: 12/16/2022] [Indexed: 12/31/2022] Open
Abstract
According to the World Health Organization statistics, as of 25 October 2022, there have been 625,248,843 confirmed cases of COVID-19, including 65,622,281 deaths worldwide. The spread and severity of COVID-19 are alarming. The economy and life of countries worldwide have been greatly affected. The rapid and accurate diagnosis of COVID-19 directly affects the spread of the virus and the degree of harm. Currently, the classification of chest X-ray or CT images based on artificial intelligence is an important method for COVID-19 diagnosis. It can assist doctors in making judgments and reduce the misdiagnosis rate. The convolutional neural network (CNN) is very popular in computer vision applications, such as applied to biological image segmentation, traffic sign recognition, face recognition, and other fields. It is one of the most widely used machine learning methods. This paper mainly introduces the latest deep learning methods and techniques for diagnosing COVID-19 using chest X-ray or CT images based on the convolutional neural network. It reviews the technology of CNN at various stages, such as rectified linear units, batch normalization, data augmentation, dropout, and so on. Several well-performing network architectures are explained in detail, such as AlexNet, ResNet, DenseNet, VGG, GoogleNet, etc. We analyzed and discussed the existing CNN automatic COVID-19 diagnosis systems from sensitivity, accuracy, precision, specificity, and F1 score. The systems use chest X-ray or CT images as datasets. Overall, CNN has essential value in COVID-19 diagnosis. All of them have good performance in the existing experiments. If expanding the datasets, adding GPU acceleration and data preprocessing techniques, and expanding the types of medical images, the performance of CNN will be further improved. This paper wishes to make contributions to future research.
Collapse
Affiliation(s)
- Xue Han
- School of Mathematics and Information Science, Nanjing Normal University of Special Education, Nanjing 210038, China
- School of Computing and Mathematical Sciences, University of Leicester, Leicester LE1 7RH, UK
| | - Zuojin Hu
- School of Mathematics and Information Science, Nanjing Normal University of Special Education, Nanjing 210038, China
| | - Shuihua Wang
- School of Computing and Mathematical Sciences, University of Leicester, Leicester LE1 7RH, UK
| | - Yudong Zhang
- School of Computing and Mathematical Sciences, University of Leicester, Leicester LE1 7RH, UK
| |
Collapse
|
11
|
Convolution-layer parameters optimization in Convolutional Neural Networks. Knowl Based Syst 2022. [DOI: 10.1016/j.knosys.2022.110210] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/23/2022]
|
12
|
Hybrid PSO (SGPSO) with the Incorporation of Discretization Operator for Training RBF Neural Network and Optimal Feature Selection. ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING 2022. [DOI: 10.1007/s13369-022-07408-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/17/2022]
|
13
|
Kaveh M, Mesgari MS. Application of Meta-Heuristic Algorithms for Training Neural Networks and Deep Learning Architectures: A Comprehensive Review. Neural Process Lett 2022; 55:1-104. [PMID: 36339645 PMCID: PMC9628382 DOI: 10.1007/s11063-022-11055-6] [Citation(s) in RCA: 9] [Impact Index Per Article: 4.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 10/11/2022] [Indexed: 12/02/2022]
Abstract
The learning process and hyper-parameter optimization of artificial neural networks (ANNs) and deep learning (DL) architectures is considered one of the most challenging machine learning problems. Several past studies have used gradient-based back propagation methods to train DL architectures. However, gradient-based methods have major drawbacks such as stucking at local minimums in multi-objective cost functions, expensive execution time due to calculating gradient information with thousands of iterations and needing the cost functions to be continuous. Since training the ANNs and DLs is an NP-hard optimization problem, their structure and parameters optimization using the meta-heuristic (MH) algorithms has been considerably raised. MH algorithms can accurately formulate the optimal estimation of DL components (such as hyper-parameter, weights, number of layers, number of neurons, learning rate, etc.). This paper provides a comprehensive review of the optimization of ANNs and DLs using MH algorithms. In this paper, we have reviewed the latest developments in the use of MH algorithms in the DL and ANN methods, presented their disadvantages and advantages, and pointed out some research directions to fill the gaps between MHs and DL methods. Moreover, it has been explained that the evolutionary hybrid architecture still has limited applicability in the literature. Also, this paper classifies the latest MH algorithms in the literature to demonstrate their effectiveness in DL and ANN training for various applications. Most researchers tend to extend novel hybrid algorithms by combining MHs to optimize the hyper-parameters of DLs and ANNs. The development of hybrid MHs helps improving algorithms performance and capable of solving complex optimization problems. In general, the optimal performance of the MHs should be able to achieve a suitable trade-off between exploration and exploitation features. Hence, this paper tries to summarize various MH algorithms in terms of the convergence trend, exploration, exploitation, and the ability to avoid local minima. The integration of MH with DLs is expected to accelerate the training process in the coming few years. However, relevant publications in this way are still rare.
Collapse
Affiliation(s)
- Mehrdad Kaveh
- Department of Geodesy and Geomatics, K. N. Toosi University of Technology, Tehran, 19967-15433 Iran
| | - Mohammad Saadi Mesgari
- Department of Geodesy and Geomatics, K. N. Toosi University of Technology, Tehran, 19967-15433 Iran
| |
Collapse
|
14
|
Wu H, Zhang X, Song L, Zhang Y, Gu L, Zhao X. Wild Geese Migration Optimization Algorithm: A New Meta-Heuristic Algorithm for Solving Inverse Kinematics of Robot. COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE 2022; 2022:5191758. [PMID: 36337271 PMCID: PMC9630037 DOI: 10.1155/2022/5191758] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 06/12/2022] [Revised: 07/21/2022] [Accepted: 08/18/2022] [Indexed: 08/26/2023]
Abstract
This paper proposes a new meta-heuristic algorithm, named wild geese migration optimization (GMO) algorithm. It is inspired by the social behavior of wild geese swarming in nature. They maintain a special formation for long-distance migration in small groups for survival and reproduction. The mathematical model is established based on these social behaviors to solve optimization problems. Meanwhile, the performance of the GMO algorithm is tested on the stable benchmark function of CEC2017, and its potential for dealing with practical problems is studied in five engineering design problems and the inverse kinematics solution of robot. The test results show that the GMO algorithm has excellent computational performance compared to other algorithms. The practical application results show that the GMO algorithm has strong applicability, more accurate optimization results, and more competitiveness in challenging problems with unknown search space, compared with well-known algorithms in the literature. The proposal of GMO algorithm enriches the team of swarm intelligence optimization algorithms and also provides a new solution for solving engineering design problems and inverse kinematics of robots.
Collapse
Affiliation(s)
- Honggang Wu
- School of Mechanical and Electrical Engineering, Changchun University of Science and Technology, Changchun 130022, China
| | - Xinming Zhang
- School of Mechanical and Electrical Engineering, Changchun University of Science and Technology, Changchun 130022, China
- School of Mechatronic Engineering and Automation, Foshan University, Foshan 528225, China
| | - Linsen Song
- School of Mechanical and Electrical Engineering, Changchun University of Science and Technology, Changchun 130022, China
| | - Yufei Zhang
- School of Computer Science and Technology, Changchun University of Science and Technology, Changchun 130022, China
| | - Lidong Gu
- School of Mechanical and Electrical Engineering, Changchun University of Science and Technology, Changchun 130022, China
| | - Xiaonan Zhao
- School of Computer Science and Technology, Changchun University of Science and Technology, Changchun 130022, China
| |
Collapse
|
15
|
Severino AGV, de Lima JMM, de Araújo FMU. Industrial Soft Sensor Optimized by Improved PSO: A Deep Representation-Learning Approach. SENSORS (BASEL, SWITZERLAND) 2022; 22:s22186887. [PMID: 36146235 PMCID: PMC9505118 DOI: 10.3390/s22186887] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 08/01/2022] [Revised: 08/14/2022] [Accepted: 08/16/2022] [Indexed: 06/07/2023]
Abstract
Soft sensors based on deep learning approaches are growing in popularity due to their ability to extract high-level features from training, improving soft sensors' performance. In the training process of such a deep model, the set of hyperparameters is critical to archive generalization and reliability. However, choosing the training hyperparameters is a complex task. Usually, a random approach defines the set of hyperparameters, which may not be adequate regarding the high number of sets and the soft sensing purposes. This work proposes the RB-PSOSAE, a Representation-Based Particle Swarm Optimization with a modified evaluation function to optimize the hyperparameter set of a Stacked AutoEncoder-based soft sensor. The evaluation function considers the mean square error (MSE) of validation and the representation of the features extracted through mutual information (MI) analysis in the pre-training step. By doing this, the RB-PSOSAE computes hyperparameters capable of supporting the training process to generate models with improved generalization and relevant hidden features. As a result, the proposed method can generate more than 16.4% improvement in RMSE compared to another standard PSO-based method and, in some cases, more than 50% improvement compared to traditional methods applied to the same real-world nonlinear industrial process. Thus, the results demonstrate better prediction performance than traditional and state-of-the-art methods.
Collapse
|
16
|
Kanipriya M, Hemalatha C, Sridevi N, SriVidhya S, Jany Shabu S. An improved capuchin search algorithm optimized hybrid CNN-LSTM architecture for malignant lung nodule detection. Biomed Signal Process Control 2022. [DOI: 10.1016/j.bspc.2022.103973] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/02/2022]
|
17
|
|
18
|
Yousif NR, Balaha HM, Haikal AY, El-Gendy EM. A generic optimization and learning framework for Parkinson disease via speech and handwritten records. JOURNAL OF AMBIENT INTELLIGENCE AND HUMANIZED COMPUTING 2022; 14:1-21. [PMID: 36042792 PMCID: PMC9411848 DOI: 10.1007/s12652-022-04342-6] [Citation(s) in RCA: 5] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 10/15/2021] [Accepted: 07/11/2022] [Indexed: 06/15/2023]
Abstract
Parkinson's disease (PD) is a neurodegenerative disorder with slow progression whose symptoms can be identified at late stages. Early diagnosis and treatment of PD can help to relieve the symptoms and delay progression. However, this is very challenging due to the similarities between the symptoms of PD and other diseases. The current study proposes a generic framework for the diagnosis of PD using handwritten images and (or) speech signals. For the handwriting images, 8 pre-trained convolutional neural networks (CNN) via transfer learning tuned by Aquila Optimizer were trained on the NewHandPD dataset to diagnose PD. For the speech signals, features from the MDVR-KCL dataset are extracted numerically using 16 feature extraction algorithms and fed to 4 different machine learning algorithms tuned by Grid Search algorithm, and graphically using 5 different techniques and fed to the 8 pretrained CNN structures. The authors propose a new technique in extracting the features from the voice dataset based on the segmentation of variable speech-signal-segment-durations, i.e., the use of different durations in the segmentation phase. Using the proposed technique, 5 datasets with 281 numerical features are generated. Results from different experiments are collected and recorded. For the NewHandPD dataset, the best-reported metric is 99.75% using the VGG19 structure. For the MDVR-KCL dataset, the best-reported metrics are 99.94% using the KNN and SVM ML algorithms and the combined numerical features; and 100% using the combined the mel-specgram graphical features and VGG19 structure. These results are better than other state-of-the-art researches.
Collapse
Affiliation(s)
- Nada R. Yousif
- Computer and Control Systems Engineering Department, Faculty of Engineering, Mansoura University, Mansoura, Egypt
| | - Hossam Magdy Balaha
- Computer and Control Systems Engineering Department, Faculty of Engineering, Mansoura University, Mansoura, Egypt
| | - Amira Y. Haikal
- Computer and Control Systems Engineering Department, Faculty of Engineering, Mansoura University, Mansoura, Egypt
| | - Eman M. El-Gendy
- Computer and Control Systems Engineering Department, Faculty of Engineering, Mansoura University, Mansoura, Egypt
| |
Collapse
|
19
|
Single candidate optimizer: a novel optimization algorithm. EVOLUTIONARY INTELLIGENCE 2022. [DOI: 10.1007/s12065-022-00762-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/15/2022]
Abstract
AbstractSingle-solution-based optimization algorithms have gained little to no attention by the research community, unlike population-based approaches. This paper proposes a novel optimization algorithm, called Single Candidate Optimizer (SCO), that relies only on a single candidate solution throughout the whole optimization process. The proposed algorithm implements a unique set of equations to effectively update the position of the candidate solution. To balance exploration and exploitation, SCO is integrated with the two-phase strategy where the candidate solution updates its position differently in each phase. The effectiveness of the proposed approach is validated by testing it on thirty three classical benchmarking functions and four real-world engineering problems. SCO is compared with three well-known optimization algorithms, i.e., Particle Swarm Optimization, Grey Wolf Optimizer, and Gravitational Search Algorithm and with four recent high-performance algorithms: Equilibrium Optimizer, Archimedes Optimization Algorithm, Mayfly Algorithm, and Salp Swarm Algorithm. According to Friedman and Wilcoxon rank-sum tests, SCO can significantly outperform all other algorithms for the majority of the investigated problems. The results achieved by SCO motivates the design and development of new single-solution-based optimization algorithms to further improve the performance. The source code of SCO is publicly available at: https://uk.mathworks.com/matlabcentral/fileexchange/116100-single-candidate-optimizer.
Collapse
|
20
|
A novel approach for optimization of convolution neural network with hybrid particle swarm and grey wolf algorithm for classification of Indian classical dances. Knowl Inf Syst 2022; 64:2411-2434. [PMID: 35919768 PMCID: PMC9333353 DOI: 10.1007/s10115-022-01707-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/12/2022] [Accepted: 06/26/2022] [Indexed: 11/02/2022]
Abstract
Deep learning is the most dominant area to perform the complex challenging tasks such as image classification and recognition. Earlier researchers have been proposed various convolution neural network (CNN) with different architectures to improve the performance accuracy for the classification and recognition of images. However, the fine-tuning of hyper parameters, resulting the optimal network, regularization of parameters is the difficult task. The metaheuristic optimization algorithms are used for solving such kind of problems. In this paper we proffer a fine tune automate CNN with Hybrid Particle Swarm Grey Wolf (HPSGW). This novel algorithm used to discover the optimal parameters of the CNN like batch size, number of hidden layers, number of epochs and size of filters. The proffered optimized architecture is implemented on MNIST, CIFAR are two bench mark datasets and Indian Classical Dance (ICD) for the classification of 8 Indian Classical Dances. The Proffered method improves the model performance accuracy of 97.3% on ICD Dataset, and other benchmark datasets MNIST, CIFAR with an improved accuracy of 99.4% and 91.1%. This auto-tuned network improved the performance by 5.6% for Indian Classical Dance Forms Classification compared to earlier methods and also reduces the computational cost.
Collapse
|
21
|
An Improved Deep Neural Network Model of Intelligent Vehicle Dynamics via Linear Decreasing Weight Particle Swarm and Invasive Weed Optimization Algorithms. SENSORS 2022; 22:s22134676. [PMID: 35808170 PMCID: PMC9269210 DOI: 10.3390/s22134676] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 05/23/2022] [Revised: 06/17/2022] [Accepted: 06/17/2022] [Indexed: 11/20/2022]
Abstract
We propose an improved DNN modeling method based on two optimization algorithms, namely the linear decreasing weight particle swarm optimization (LDWPSO) algorithm and invasive weed optimization (IWO) algorithm, for predicting vehicle’s longitudinal-lateral responses. The proposed improved method can restrain the solutions of weight matrices and bias matrices from falling into a local optimum while training the DNN model. First, dynamic simulations for a vehicle are performed based on an efficient semirecursive multibody model for real-time data acquisition. Next, the vehicle data are processed and used to train and test the improved DNN model. The vehicle responses, which are obtained from the LDWPSO-DNN and IWO-DNN models, are compared with the DNN and multibody results. The comparative results show that the LDWPSO-DNN and IWO-DNN models predict accurate longitudinal-lateral responses in real-time without falling into a local optimum. The improved DNN model based on optimization algorithms can be employed for real-time simulation and preview control in intelligent vehicles.
Collapse
|
22
|
Wang Y, Li J, Chen C, Zhang J, Zhan Z. Scale adaptive fitness evaluation‐based particle swarm optimisation for hyperparameter and architecture optimisation in neural networks and deep learning. CAAI TRANSACTIONS ON INTELLIGENCE TECHNOLOGY 2022. [DOI: 10.1049/cit2.12106] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/19/2022] Open
Affiliation(s)
- Ye‐Qun Wang
- School of Computer Science and Engineering South China University of Technology Guangzhou China
| | - Jian‐Yu Li
- School of Computer Science and Engineering South China University of Technology Guangzhou China
| | - Chun‐Hua Chen
- School of Software Engineering South China University of Technology Guangzhou China
| | - Jun Zhang
- Hanyang University Ansan South Korea
| | - Zhi‐Hui Zhan
- School of Computer Science and Engineering South China University of Technology Guangzhou China
| |
Collapse
|
23
|
Abstract
Convolutional Neural Networks (CNNs) operate within a wide variety of hyperparameters, the optimization of which can greatly improve the performance of CNNs when performing the task at hand. However, these hyperparameters can be very difficult to optimize, either manually or by brute force. Neural architecture search or NAS methods have been developed to address this problem and are used to find the best architectures for the deep learning paradigm. In this article, a CNN has been evolved with a well-known nature-inspired metaheuristic paddy field algorithm (PFA). It can be seen that PFA can evolve the neural architecture using the Google Landmarks Dataset V2, which is one of the toughest datasets available in the literature. The CNN’s performance, when evaluated based on the accuracy benchmark, increases from an accuracy of 0.53 to 0.76, which is an improvement of more than 40%. The evolved architecture also shows some major improvements in hyperparameters that are normally considered to be the best suited for the task.
Collapse
|
24
|
Slade S, Zhang L, Yu Y, Lim CP. An evolving ensemble model of multi-stream convolutional neural networks for human action recognition in still images. Neural Comput Appl 2022. [DOI: 10.1007/s00521-022-06947-6] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/19/2022]
Abstract
AbstractStill image human action recognition (HAR) is a challenging problem owing to limited sources of information and large intra-class and small inter-class variations which requires highly discriminative features. Transfer learning offers the necessary capabilities in producing such features by preserving prior knowledge while learning new representations. However, optimally identifying dynamic numbers of re-trainable layers in the transfer learning process poses a challenge. In this study, we aim to automate the process of optimal configuration identification. Specifically, we propose a novel particle swarm optimisation (PSO) variant, denoted as EnvPSO, for optimal hyper-parameter selection in the transfer learning process with respect to HAR tasks with still images. It incorporates Gaussian fitness surface prediction and exponential search coefficients to overcome stagnation. It optimises the learning rate, batch size, and number of re-trained layers of a pre-trained convolutional neural network (CNN). To overcome bias of single optimised networks, an ensemble model with three optimised CNN streams is introduced. The first and second streams employ raw images and segmentation masks yielded by mask R-CNN as inputs, while the third stream fuses a pair of networks with raw image and saliency maps as inputs, respectively. The final prediction results are obtained by computing the average of class predictions from all three streams. By leveraging differences between learned representations within optimised streams, our ensemble model outperforms counterparts devised by PSO and other state-of-the-art methods for HAR. In addition, evaluated using diverse artificial landscape functions, EnvPSO performs better than other search methods with statistically significant difference in performance.
Collapse
|
25
|
Recognition of shed damage on 11-kV polymer insulator using Bayesian optimized convolution neural network. Soft comput 2022. [DOI: 10.1007/s00500-021-06629-w] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/19/2022]
|
26
|
Layek K, Basak B, Samanta S, Maity SP, Barui A. Stiffness prediction on elastography images and neuro-fuzzy based segmentation for thyroid cancer detection. APPLIED OPTICS 2022; 61:49-59. [PMID: 35200805 DOI: 10.1364/ao.445226] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 10/05/2021] [Accepted: 11/29/2021] [Indexed: 06/14/2023]
Abstract
The elastography method detects metastatic changes by measuring the stiffness of tissues. Estimation of elasticities from elastography images facilitates more precise identification of the metastatic region and detection of the same. In this study, an automated segmentation algorithm is proposed that calculates pixel-wise elasticity values to detect thyroid cancer from elastography images. This intensity to elasticity conversion is achieved by constructing a fuzzy inference system using an adaptive neuro-fuzzy inference system supported by two meta-heuristic algorithms: genetic algorithm and particle swarm optimization. Pixels of the input color images (red, green, and blue) are replaced by equivalent elasticity values (in kilo Pascal) and are stored in a two-dimensional array to form an "elasticity matrix." The elasticity matrix is then segmented into three regions, namely, suspicious, near-suspicious, and non-suspicious, based on the elasticity measures, where the threshold limits are calculated using the fuzzy entropy maximization method optimized by the differential evolution algorithm. Segmentation performances are evaluated by Kappa and the dice similarity co-efficient, and average values achieved are 0.94±0.11 and 0.93±0.12, respectively. Sensitivity and specificity values achieved by the proposed method are 86.35±0.34% and 97.67±0.40%, respectively, showing an overall accuracy of 93.50±0.42%. Results justify the importance of pixel stiffness for segmentation of thyroid nodules in elastography images.
Collapse
|
27
|
Yogarathinam LT, Velswamy K, Gangasalam A, Ismail AF, Goh PS, Narayanan A, Abdullah MS. Performance evaluation of whey flux in dead-end and cross-flow modes via convolutional neural networks. JOURNAL OF ENVIRONMENTAL MANAGEMENT 2022; 301:113872. [PMID: 34607142 DOI: 10.1016/j.jenvman.2021.113872] [Citation(s) in RCA: 6] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 01/27/2021] [Revised: 09/08/2021] [Accepted: 09/26/2021] [Indexed: 06/13/2023]
Abstract
Effluent originating from cheese production puts pressure onto environment due to its high organic load. Therefore, the main objective of this work was to compare the influence of different process variables (transmembrane pressure (TMP), Reynolds number and feed pH) on whey protein recovery from synthetic and industrial cheese whey using polyethersulfone (PES 30 kDa) membrane in dead-end and cross-flow modes. Analysis on the fouling mechanistic model indicates that cake layer formation is dominant as compared to other pore blocking phenomena evaluated. Among the input variables, pH of whey protein solution has the biggest influence towards membrane flux and protein rejection performances. At pH 4, electrostatic attraction experienced by whey protein molecules prompted a decline in flux. Cross-flow filtration system exhibited a whey rejection value of 0.97 with an average flux of 69.40 L/m2h and at an experimental condition of 250 kPa and 8 for TMP and pH, respectively. The dynamic behavior of whey effluent flux was modeled using machine learning (ML) tool convolutional neural networks (CNN) and recursive one-step prediction scheme was utilized. Linear and non-linear correlation indicated that CNN model (R2 - 0.99) correlated well with the dynamic flux experimental data. PES 30 kDa membrane displayed a total protein rejection coefficient of 0.96 with 55% of water recovery for the industrial cheese whey effluent. Overall, these filtration studies revealed that this dynamic whey flux data studies using the CNN modeling also has a wider scope as it can be applied in sensor tuning to monitor flux online by means of enhancing whey recovery efficiency.
Collapse
Affiliation(s)
- Lukka Thuyavan Yogarathinam
- Membrane Research Laboratory, Department of Chemical Engineering, National Institute of Technology, Tiruchirappalli, 620 015, India; Advanced Membrane Technology Research Centre (AMTEC), Universiti Teknologi Malaysia, 81310, Skudai, Johor, Malaysia
| | - Kirubakaran Velswamy
- Department of Chemical and Materials Engineering, Donadeo Innovation Center for Engineering, University of Alberta-T6G 1H9, Edmonton, Canada
| | - Arthanareeswaran Gangasalam
- Membrane Research Laboratory, Department of Chemical Engineering, National Institute of Technology, Tiruchirappalli, 620 015, India.
| | - Ahmad Fauzi Ismail
- Advanced Membrane Technology Research Centre (AMTEC), Universiti Teknologi Malaysia, 81310, Skudai, Johor, Malaysia.
| | - Pei Sean Goh
- Advanced Membrane Technology Research Centre (AMTEC), Universiti Teknologi Malaysia, 81310, Skudai, Johor, Malaysia
| | - Anantharaman Narayanan
- Membrane Research Laboratory, Department of Chemical Engineering, National Institute of Technology, Tiruchirappalli, 620 015, India
| | - Mohd Sohaimi Abdullah
- Advanced Membrane Technology Research Centre (AMTEC), Universiti Teknologi Malaysia, 81310, Skudai, Johor, Malaysia
| |
Collapse
|
28
|
Miao F, Yao L, Zhao X. Evolving convolutional neural networks by symbiotic organisms search algorithm for image classification. Appl Soft Comput 2021. [DOI: 10.1016/j.asoc.2021.107537] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/28/2022]
|
29
|
Zhou S, Han Y, Sha L, Zhu S. A multi-sample particle swarm optimization algorithm based on electric field force. MATHEMATICAL BIOSCIENCES AND ENGINEERING : MBE 2021; 18:7464-7489. [PMID: 34814258 DOI: 10.3934/mbe.2021369] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/13/2023]
Abstract
Aiming at the premature convergence problem of particle swarm optimization algorithm, a multi-sample particle swarm optimization (MSPSO) algorithm based on electric field force is proposed. Firstly, we introduce the concept of the electric field into the particle swarm optimization algorithm. The particles are affected by the electric field force, which makes the particles exhibit diverse behaviors. Secondly, MSPSO constructs multiple samples through two new strategies to guide particle learning. An electric field force-based comprehensive learning strategy (EFCLS) is proposed to build attractive samples and repulsive samples, thus improving search efficiency. To further enhance the convergence accuracy of the algorithm, a segment-based weighted learning strategy (SWLS) is employed to construct a global learning sample so that the particles learn more comprehensive information. In addition, the parameters of the model are adjusted adaptively to adapt to the population status in different periods. We have verified the effectiveness of these newly proposed strategies through experiments. Sixteen benchmark functions and eight well-known particle swarm optimization algorithm variants are employed to prove the superiority of MSPSO. The comparison results show that MSPSO has better performance in terms of accuracy, especially for high-dimensional spaces, while maintaining a faster convergence rate. Besides, a real-world problem also verified that MSPSO has practical application value.
Collapse
Affiliation(s)
- Shangbo Zhou
- College of Computer Science, Chongqing University, Chongqing 400044, China
- Key Laboratory of Dependable Service Computing in Cyber Physical Society, Ministry of Education, Chongqing 400030, China
| | - Yuxiao Han
- College of Computer Science, Chongqing University, Chongqing 400044, China
- Key Laboratory of Dependable Service Computing in Cyber Physical Society, Ministry of Education, Chongqing 400030, China
| | - Long Sha
- College of Computer Science, Chongqing University, Chongqing 400044, China
- Key Laboratory of Dependable Service Computing in Cyber Physical Society, Ministry of Education, Chongqing 400030, China
| | - Shufang Zhu
- College of Computer Science, Chongqing University, Chongqing 400044, China
- Key Laboratory of Dependable Service Computing in Cyber Physical Society, Ministry of Education, Chongqing 400030, China
| |
Collapse
|
30
|
Optimization of Convolutional Neural Networks Architectures Using PSO for Sign Language Recognition. AXIOMS 2021. [DOI: 10.3390/axioms10030139] [Citation(s) in RCA: 7] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 01/02/2023]
Abstract
This paper presents an approach to design convolutional neural network architectures, using the particle swarm optimization algorithm. The adjustment of the hyper-parameters and finding the optimal network architecture of convolutional neural networks represents an important challenge. Network performance and achieving efficient learning models for a particular problem depends on setting hyper-parameter values and this implies exploring a huge and complex search space. The use of heuristic-based searches supports these types of problems; therefore, the main contribution of this research work is to apply the PSO algorithm to find the optimal parameters of the convolutional neural networks which include the number of convolutional layers, the filter size used in the convolutional process, the number of convolutional filters, and the batch size. This work describes two optimization approaches; the first, the parameters obtained by PSO are kept under the same conditions in each convolutional layer, and the objective function evaluated by PSO is given by the classification rate; in the second, the PSO generates different parameters per layer, and the objective function is composed of the recognition rate in conjunction with the Akaike information criterion, the latter helps to find the best network performance but with the minimum parameters. The optimized architectures are implemented in three study cases of sign language databases, in which are included the Mexican Sign Language alphabet, the American Sign Language MNIST, and the American Sign Language alphabet. According to the results, the proposed methodologies achieved favorable results with a recognition rate higher than 99%, showing competitive results compared to other state-of-the-art approaches.
Collapse
|