Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003080
Md Meftahul Ferdaus, S. Anavatti, M. Garratt, Mahardhika Pratama
In this paper, an adaptive Takagi-Sugeno (TS)-fuzzy controller is developed for nonlinear dynamical systems, where a new structure of the controller with reduced learning parameters is proposed. The proposed controller is named as a reduced learning parameter based fuzzy logic controller (Red-FLC). Being a model-free controller, the classical TS-fuzzy one performs well in slow-process control-based complex applications. However, the controller’s structure is associated with several antecedent and consequent parameters, which need to be adapted during control operation. Adaptation of a high number of parameters is computationally expensive, especially in controlling a system where a fast response is expected. From this research gap, in our developed adaptive fuzzy controller, the tuning parameters have reduced significantly since it has no antecedent parameters. The closed-loop stability of the controller has been proved using a new adaptation law. To evaluate the proposed controller’s performance, it has been utilized to stabilize an inverted pendulum’s simulated plant on a cart by considering an impulse disturbance. The performance of Red-FLC has been compared with a classical TS-fuzzy controller and a Proportional Integral Derivative (PID) controller, where better tracking of the cart’s position and better disturbance rejection is observed from the proposed TS-fuzzy controller.
{"title":"Red-FLC: an Adaptive Fuzzy Logic Controller with Reduced Learning Parameters","authors":"Md Meftahul Ferdaus, S. Anavatti, M. Garratt, Mahardhika Pratama","doi":"10.1109/SSCI44817.2019.9003080","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003080","url":null,"abstract":"In this paper, an adaptive Takagi-Sugeno (TS)-fuzzy controller is developed for nonlinear dynamical systems, where a new structure of the controller with reduced learning parameters is proposed. The proposed controller is named as a reduced learning parameter based fuzzy logic controller (Red-FLC). Being a model-free controller, the classical TS-fuzzy one performs well in slow-process control-based complex applications. However, the controller’s structure is associated with several antecedent and consequent parameters, which need to be adapted during control operation. Adaptation of a high number of parameters is computationally expensive, especially in controlling a system where a fast response is expected. From this research gap, in our developed adaptive fuzzy controller, the tuning parameters have reduced significantly since it has no antecedent parameters. The closed-loop stability of the controller has been proved using a new adaptation law. To evaluate the proposed controller’s performance, it has been utilized to stabilize an inverted pendulum’s simulated plant on a cart by considering an impulse disturbance. The performance of Red-FLC has been compared with a classical TS-fuzzy controller and a Proportional Integral Derivative (PID) controller, where better tracking of the cart’s position and better disturbance rejection is observed from the proposed TS-fuzzy controller.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"35 1","pages":"513-518"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77746022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the development of modern medicine, more and more methods can be used to diagnose cardiac diseases. At present, the conventional detection methods include X-ray, computed tomography (CT), magnetic resonance imaging (MRI), B-ultrasound, etc. However, the detection results are usually not intuitive enough, and it is difficult to support the sequential operation. We designed a multi-functional system based on the virtual reality (VR) equipment. We use CT and myocardial perfusion imaging (MPI) as data sources to build accurate cardiac models, and users can make operations such as rotation and scaling on these models. Experiment results have shown that the system allows users to observe cardiac tissue more intuitively, and realizes functions for the surgical program. This enables users to understand the cardiac condition more intuitively and accurately.
{"title":"A Virtual Reality System for Accurate Cardiac Modeling and Multiple Virtual Operations","authors":"Xiumei Cai, Puze Cheng, Hao-yang Shi, Cong Guo, Shao-jie Tang, Mingle Qiu","doi":"10.1109/SSCI44817.2019.9003110","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003110","url":null,"abstract":"With the development of modern medicine, more and more methods can be used to diagnose cardiac diseases. At present, the conventional detection methods include X-ray, computed tomography (CT), magnetic resonance imaging (MRI), B-ultrasound, etc. However, the detection results are usually not intuitive enough, and it is difficult to support the sequential operation. We designed a multi-functional system based on the virtual reality (VR) equipment. We use CT and myocardial perfusion imaging (MPI) as data sources to build accurate cardiac models, and users can make operations such as rotation and scaling on these models. Experiment results have shown that the system allows users to observe cardiac tissue more intuitively, and realizes functions for the surgical program. This enables users to understand the cardiac condition more intuitively and accurately.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"32 1","pages":"2782-2787"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81582438","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003001
Qing Wu, Haoyi Zhang, Rongrong Jing, Yiran Li
Twin support vector regression (TSVR) is a regression algorithm based on the support vector regression (SVR) and the spirit of the support vector machine (TWSVM) . However, some feature selection algorithms of support vector regression, such as recursive feature elimination, can’t be applied to TSVR, so a recursive feature selection method based on TSVR is proposed. By analyzing the weights, the ε -insensitive upper and lower bound functions in TSVR are analyzed. The two weight vectors are merged, and the weight vector is sorted and deleted with reference to the recursive feature elimination (RFE). The experimental results on several UCI datasets demonstrate demonstrate the effectiveness of the algorithm on feature selection and improves the regression performance.
{"title":"Feature Selection Based on Twin Support Vector Regression","authors":"Qing Wu, Haoyi Zhang, Rongrong Jing, Yiran Li","doi":"10.1109/SSCI44817.2019.9003001","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003001","url":null,"abstract":"Twin support vector regression (TSVR) is a regression algorithm based on the support vector regression (SVR) and the spirit of the support vector machine (TWSVM) . However, some feature selection algorithms of support vector regression, such as recursive feature elimination, can’t be applied to TSVR, so a recursive feature selection method based on TSVR is proposed. By analyzing the weights, the ε -insensitive upper and lower bound functions in TSVR are analyzed. The two weight vectors are merged, and the weight vector is sorted and deleted with reference to the recursive feature elimination (RFE). The experimental results on several UCI datasets demonstrate demonstrate the effectiveness of the algorithm on feature selection and improves the regression performance.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"26 1","pages":"2903-2907"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81882778","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002810
Chuang Liu, Wanghui Shen, Yingkui Du, Jiahao Lei, Ao Li
In this paper, an optimization algorithm based on membrane system is proposed for numerical optimization problems. In the proposed algorithm, we designed two mechanisms to simulate the movement of molecules in arbitrary direction and a certain direction to balance global exploration and local exploitation. To test the performance of the proposed algorithm, eight benchmark functions were chosen. The simulation results show that the proposed algorithm is more advantageous than other experimental algorithms in solving numerical optimization problems.
{"title":"Membrane System-based Optimization Algorithm for Numeric Optimization Problem","authors":"Chuang Liu, Wanghui Shen, Yingkui Du, Jiahao Lei, Ao Li","doi":"10.1109/SSCI44817.2019.9002810","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002810","url":null,"abstract":"In this paper, an optimization algorithm based on membrane system is proposed for numerical optimization problems. In the proposed algorithm, we designed two mechanisms to simulate the movement of molecules in arbitrary direction and a certain direction to balance global exploration and local exploitation. To test the performance of the proposed algorithm, eight benchmark functions were chosen. The simulation results show that the proposed algorithm is more advantageous than other experimental algorithms in solving numerical optimization problems.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"12 1","pages":"1041-1047"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84523264","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Firefighters are primarily tasked to handle fire incidents, but they are often exposed to high risks when extinguishing fire. Firefighting robots are actively being researched to reduce fire fighters injuries and deaths as well as increase their effectiveness on performing tasks. However, the major concern is how to make the flame detection methods to satisfy the high precision requirement of firefighting robot. Therefore, the flame detection of firefighting robot has become a hot topic in this area. In this paper, a Faster R-CNN model is proposed to detect flame in noisy images of fire ground. Firstly, the region generation network is used to extract the candidate flame regions. Secondly, the candidate flame regions are convoluted and pooled to extract the flame characteristics. Thirdly, the output features of Region Proposal Network (RPN) are fed into two fully connected layers: a box-regression layer which recognizes the locations of objects and a box-classification layer which classifies the objects. The dataset used in the experiment was obtained by video capture. The network is pre-trained based on Google platform Tensorflow, and the obtained precision and frame rate of the proposed method are up to 99.8% and 1.4 FPS, respectively. The experimental results demonstrate that the method equipped merits such as automatically extract the flame characteristics, effectively improve the precision of flame detection, and has excellent generalization ability and robustness.
{"title":"Faster R-CNN Based Indoor Flame Detection for Firefighting Robot","authors":"Jiadong Guo, Zengguang Hou, Xiaoliang Xie, Shuncai Yao, Qiaoli Wang, Xuechen Jin","doi":"10.1109/SSCI44817.2019.9002843","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002843","url":null,"abstract":"Firefighters are primarily tasked to handle fire incidents, but they are often exposed to high risks when extinguishing fire. Firefighting robots are actively being researched to reduce fire fighters injuries and deaths as well as increase their effectiveness on performing tasks. However, the major concern is how to make the flame detection methods to satisfy the high precision requirement of firefighting robot. Therefore, the flame detection of firefighting robot has become a hot topic in this area. In this paper, a Faster R-CNN model is proposed to detect flame in noisy images of fire ground. Firstly, the region generation network is used to extract the candidate flame regions. Secondly, the candidate flame regions are convoluted and pooled to extract the flame characteristics. Thirdly, the output features of Region Proposal Network (RPN) are fed into two fully connected layers: a box-regression layer which recognizes the locations of objects and a box-classification layer which classifies the objects. The dataset used in the experiment was obtained by video capture. The network is pre-trained based on Google platform Tensorflow, and the obtained precision and frame rate of the proposed method are up to 99.8% and 1.4 FPS, respectively. The experimental results demonstrate that the method equipped merits such as automatically extract the flame characteristics, effectively improve the precision of flame detection, and has excellent generalization ability and robustness.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"39 1","pages":"1390-1395"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85033793","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003053
Josiah Laivins, Minwoo Lee
Even with recent advances in standard reinforcement learning, hierarchical reinforcement learning has been discussed as a promising approach to solve complex problems. From human-designed abstraction, planning or learning with composite actions are well-understood, but without human intervention, producing abstract (or composite) actions automatically is one of the remaining challenges. We separate this action discovery from reinforcement learning problem and investigate on searching impactful composite actions that can make meaningful changes in state space. We discuss the efficiency and flexibility of the suggested model by interpreting and analyzing the discovered composite actions with different deep reinforcement learning algorithms in different environments.
{"title":"Automatic Composite Action Discovery for Hierarchical Reinforcement Learning","authors":"Josiah Laivins, Minwoo Lee","doi":"10.1109/SSCI44817.2019.9003053","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003053","url":null,"abstract":"Even with recent advances in standard reinforcement learning, hierarchical reinforcement learning has been discussed as a promising approach to solve complex problems. From human-designed abstraction, planning or learning with composite actions are well-understood, but without human intervention, producing abstract (or composite) actions automatically is one of the remaining challenges. We separate this action discovery from reinforcement learning problem and investigate on searching impactful composite actions that can make meaningful changes in state space. We discuss the efficiency and flexibility of the suggested model by interpreting and analyzing the discovered composite actions with different deep reinforcement learning algorithms in different environments.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"56 1","pages":"198-205"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85882173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002928
Chunjie Yang, Bingchun Jiao, Hongbo Kang, Yanwei Li, Yan Liu, Yifan Wu, Shujie Ma
At present, wired method is used widely in the acquisition and transmission of field signals in industrial field acquisition system. There are some problems such as long cables, difficult wiring, analog signals that are susceptible to interference and so on. A new design of industrial field intelligent temperature acquisition system based on timestamped anti-interference algorithm is presented in the paper. The system can collect analog signals of PT100 and convert that to digital signals which are transmitted to the control center by Data Transceiver. When the signals reach the control center, the system will process digital signals, and set up T-R relation table based on least square algorithm, and at last, the signals are restored to resistance signals of PT100 by using multichannel digital potentiometers. The system uses wireless transmission instead of wired transmission, to realize the collection, transmission and recovery of thermal resistance information, and solve the interference problem of long-distance wireless communication. Experimental test verify that the system communication is normal, stability and accurate, which provides a reasonable design scheme for the application of data transmission technology in industrial field.
{"title":"Design of Industrial Field Intelligent Temperature Acquisition System Based on Timestamped Anti-Interference Algorithm","authors":"Chunjie Yang, Bingchun Jiao, Hongbo Kang, Yanwei Li, Yan Liu, Yifan Wu, Shujie Ma","doi":"10.1109/SSCI44817.2019.9002928","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002928","url":null,"abstract":"At present, wired method is used widely in the acquisition and transmission of field signals in industrial field acquisition system. There are some problems such as long cables, difficult wiring, analog signals that are susceptible to interference and so on. A new design of industrial field intelligent temperature acquisition system based on timestamped anti-interference algorithm is presented in the paper. The system can collect analog signals of PT100 and convert that to digital signals which are transmitted to the control center by Data Transceiver. When the signals reach the control center, the system will process digital signals, and set up T-R relation table based on least square algorithm, and at last, the signals are restored to resistance signals of PT100 by using multichannel digital potentiometers. The system uses wireless transmission instead of wired transmission, to realize the collection, transmission and recovery of thermal resistance information, and solve the interference problem of long-distance wireless communication. Experimental test verify that the system communication is normal, stability and accurate, which provides a reasonable design scheme for the application of data transmission technology in industrial field.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"31 1","pages":"2884-2889"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85916987","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003037
Hanxiao Qian, Pengjie Gu, Rui Yan, Huajin Tang
In this paper, we propose a robust multi-label classification system based on deep spiking neural networks to handle multi-pitch estimation tasks. We employ constantQ transform spectrogram as a time-frequency representation. A keypoint detection technique is used for noise suppression and the extraction of relevant information. We also propose a novel biological spiking coding method that fits the expression of musical signals. This coding method can encode time, frequency, intensity information into spatiotemporal spike trains. And the spatio-temporal credit assignment (STCA) algorithm is used to train deep spiking neural networks. We perform the multipitch evaluation on the MAPS data set, and our work compares with the state-of-the-art methods by using the F1-score metric. Experimental results show that the proposed scheme has achieved better performance than other state-of-the-art methods and reveal the system’s robustness to environmental noise.
{"title":"Robust Multipitch Estimation of Piano Sounds Using Deep Spiking Neural Networks","authors":"Hanxiao Qian, Pengjie Gu, Rui Yan, Huajin Tang","doi":"10.1109/SSCI44817.2019.9003037","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003037","url":null,"abstract":"In this paper, we propose a robust multi-label classification system based on deep spiking neural networks to handle multi-pitch estimation tasks. We employ constantQ transform spectrogram as a time-frequency representation. A keypoint detection technique is used for noise suppression and the extraction of relevant information. We also propose a novel biological spiking coding method that fits the expression of musical signals. This coding method can encode time, frequency, intensity information into spatiotemporal spike trains. And the spatio-temporal credit assignment (STCA) algorithm is used to train deep spiking neural networks. We perform the multipitch evaluation on the MAPS data set, and our work compares with the state-of-the-art methods by using the F1-score metric. Experimental results show that the proposed scheme has achieved better performance than other state-of-the-art methods and reveal the system’s robustness to environmental noise.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"6 1","pages":"2335-2341"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80854090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9003134
Duc C. Le, A. N. Zincir-Heywood, M. Heywood
Different variations in deployment environments of machine learning techniques may affect the performance of the implemented systems. The variations may cause changes in the data for machine learning solutions, such as in the number of classes and the extracted features. This paper investigates the capabilities of Genetic Programming (GP) for malicious insider detection in corporate environments under such changes. Assuming a Linear GP detector, techniques are introduced to allow a previously trained GP population to adapt to different changes in the data. The experiments and evaluation results show promising insider threat detection performances of the techniques in comparison with training machine learning classifiers from scratch. This reduces the amount of data needed and computation requirements for obtaining dependable insider threat detectors under new conditions.
{"title":"Dynamic Insider Threat Detection Based on Adaptable Genetic Programming","authors":"Duc C. Le, A. N. Zincir-Heywood, M. Heywood","doi":"10.1109/SSCI44817.2019.9003134","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9003134","url":null,"abstract":"Different variations in deployment environments of machine learning techniques may affect the performance of the implemented systems. The variations may cause changes in the data for machine learning solutions, such as in the number of classes and the extracted features. This paper investigates the capabilities of Genetic Programming (GP) for malicious insider detection in corporate environments under such changes. Assuming a Linear GP detector, techniques are introduced to allow a previously trained GP population to adapt to different changes in the data. The experiments and evaluation results show promising insider threat detection performances of the techniques in comparison with training machine learning classifiers from scratch. This reduces the amount of data needed and computation requirements for obtaining dependable insider threat detectors under new conditions.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"26 1","pages":"2579-2586"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81242531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-12-01DOI: 10.1109/SSCI44817.2019.9002978
Chengcheng Liu, Huikai Shao, Dexing Zhong, Jun Du
In recent years, palmprint-based recognition technology has become one of the hotspots in biometrics research. The accuracy of traditional palmprint recognition algorithms mainly depends on vast data and labels. However, in reality, we usually have few labeled data. To solve this problem, the paper explores the application of few-shot recognition to palmprint. In the preprocessing stage, a novel region of interest (ROI) extraction algorithm is proposed, which can extract more palmprint texture features in the relatively fixed palm area and effectively improve the impact of palm size on preprocessing results. In the feature extraction stage, the paper presents a nonpooling Siamese-Hashing Network structure, called SHN. This method can extract high discriminant features of new categories from only a small number of samples. In addition, the output of SHN is a 48-bit hashing code, which takes up less memory and matches samples faster. Experiment results show that the performance of the model in the benchmark database is better than other classical models in the few-shot case.
{"title":"Siamese-Hashing Network for Few-Shot Palmprint Recognition","authors":"Chengcheng Liu, Huikai Shao, Dexing Zhong, Jun Du","doi":"10.1109/SSCI44817.2019.9002978","DOIUrl":"https://doi.org/10.1109/SSCI44817.2019.9002978","url":null,"abstract":"In recent years, palmprint-based recognition technology has become one of the hotspots in biometrics research. The accuracy of traditional palmprint recognition algorithms mainly depends on vast data and labels. However, in reality, we usually have few labeled data. To solve this problem, the paper explores the application of few-shot recognition to palmprint. In the preprocessing stage, a novel region of interest (ROI) extraction algorithm is proposed, which can extract more palmprint texture features in the relatively fixed palm area and effectively improve the impact of palm size on preprocessing results. In the feature extraction stage, the paper presents a nonpooling Siamese-Hashing Network structure, called SHN. This method can extract high discriminant features of new categories from only a small number of samples. In addition, the output of SHN is a 48-bit hashing code, which takes up less memory and matches samples faster. Experiment results show that the performance of the model in the benchmark database is better than other classical models in the few-shot case.","PeriodicalId":6729,"journal":{"name":"2019 IEEE Symposium Series on Computational Intelligence (SSCI)","volume":"11 1","pages":"3251-3258"},"PeriodicalIF":0.0,"publicationDate":"2019-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79516287","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：