Pub Date : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642182
Kung-Ting Wei, Yaojun Chu, Haiyun Gan
Aiming at solving the issue that the existing Rapidly-exploring Random Tree (RRT) algorithm cannot well replan the paths to avoid dynamic obstacles for robotic manipulator autonomously and rapidly in complex cluttered environments, three-dimensional reconstruction of the global dynamic scene around the robotic manipulator is carried out based on RGB-D visual sensor in this paper. A Bi-RRT-Star dynamic path planning approach based on improved exploring function with goal direction is proposed, which is improved from connection strategy, heuristic intensive exploring, and adjacent nodes expansion. On this basis, a multi-step expansion strategy with heuristic greedy is presented. Finally, the relevant evaluation indices of the proposed approach are verified in Virtual Robot Environment Platform (VREP) software. The simulation results show that in comparison with Bi-RRT and RRT-Star algorithms, the proposed method has a higher success rate in dynamic path planning online with less planning time and lower trajectory cost. In addition, a realistic experiment is designed to make UR robotic manipulator avoid human arm random motions dynamically. The experimental results show that the proposed method successfully realizes that robotic manipulator can avoid continuous moving obstacles of human arm online smoothly, comprehensively verifying the effectiveness and superiority.
针对现有快速探索随机树(rapid -exploring Random Tree, RRT)算法在复杂杂乱环境中无法自主快速地重新规划机械臂避开动态障碍物的路径的问题，本文基于RGB-D视觉传感器对机械臂周围全局动态场景进行了三维重建。从连接策略、启发式密集探索和相邻节点扩展三个方面进行改进，提出了一种基于改进的带目标方向探索函数的Bi-RRT-Star动态路径规划方法。在此基础上，提出了一种带有启发式贪婪的多步展开策略。最后，在虚拟机器人环境平台(VREP)软件中对所提方法的相关评价指标进行了验证。仿真结果表明，与Bi-RRT和RRT-Star算法相比，该方法具有较高的在线动态路径规划成功率，规划时间短，轨迹成本低。此外，还设计了一个现实实验，使UR机械臂能够动态地避免人臂的随机运动。实验结果表明，所提方法成功地实现了机械臂能够在线顺利避开人臂连续移动障碍物，全面验证了该方法的有效性和优越性。
{"title":"An improved Rapidly-exploring Random Tree Approach for Robotic Dynamic Path Planning","authors":"Kung-Ting Wei, Yaojun Chu, Haiyun Gan","doi":"10.1109/ICICIP53388.2021.9642182","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642182","url":null,"abstract":"Aiming at solving the issue that the existing Rapidly-exploring Random Tree (RRT) algorithm cannot well replan the paths to avoid dynamic obstacles for robotic manipulator autonomously and rapidly in complex cluttered environments, three-dimensional reconstruction of the global dynamic scene around the robotic manipulator is carried out based on RGB-D visual sensor in this paper. A Bi-RRT-Star dynamic path planning approach based on improved exploring function with goal direction is proposed, which is improved from connection strategy, heuristic intensive exploring, and adjacent nodes expansion. On this basis, a multi-step expansion strategy with heuristic greedy is presented. Finally, the relevant evaluation indices of the proposed approach are verified in Virtual Robot Environment Platform (VREP) software. The simulation results show that in comparison with Bi-RRT and RRT-Star algorithms, the proposed method has a higher success rate in dynamic path planning online with less planning time and lower trajectory cost. In addition, a realistic experiment is designed to make UR robotic manipulator avoid human arm random motions dynamically. The experimental results show that the proposed method successfully realizes that robotic manipulator can avoid continuous moving obstacles of human arm online smoothly, comprehensively verifying the effectiveness and superiority.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127978036","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642210
Shiyuan Fang, Hanzhi Li, Dacheng Pei, Meixi Wu, Yichong Sun, B. Cai
This paper concerns the problem of stochastic stability and stabilization for a class of discrete-time hidden semi-Markov jump systems (HS-MJSs) with time-varying emission probability. By virtue of the hidden semi-Markov model, such stochastic switching system has shown superior capacity of describing the asynchronous phenomenon between the practical system mode and the observed one. Moreover, the proposed emission probability is time-varying and satisfies piecewise homogeneous property in this paper. Based on the presented Lyapunov function, an observed-mode-dependent controller is constructed to stabilize the closed-loop HS-MJSs. Finally, the effectiveness of the proposed control scheme has been verified by a numerical example.
{"title":"Stabilization of Discrete-Time Hidden Semi-Markov Jump Systems with Time-varying Emission Probability","authors":"Shiyuan Fang, Hanzhi Li, Dacheng Pei, Meixi Wu, Yichong Sun, B. Cai","doi":"10.1109/ICICIP53388.2021.9642210","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642210","url":null,"abstract":"This paper concerns the problem of stochastic stability and stabilization for a class of discrete-time hidden semi-Markov jump systems (HS-MJSs) with time-varying emission probability. By virtue of the hidden semi-Markov model, such stochastic switching system has shown superior capacity of describing the asynchronous phenomenon between the practical system mode and the observed one. Moreover, the proposed emission probability is time-varying and satisfies piecewise homogeneous property in this paper. Based on the presented Lyapunov function, an observed-mode-dependent controller is constructed to stabilize the closed-loop HS-MJSs. Finally, the effectiveness of the proposed control scheme has been verified by a numerical example.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129867374","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642176
Xinhui Zhu, Li Zhang, Yang Shi, Jing Wang, Jian Li
Robot manipulator control is a complicated multi-tasking problem in reality. It includes not only basic tracking task, but also additional tasks, such as conquering joint angle limits, posture control. However, most existing works only consider the goal of tracking and formulate it as single-layered time-variant problems, which leads to impracticality. In this work, robot manipulator control problem is formulated as four-layered time-variant equations including linear, nonlinear equalities and inequalities. Each layer formulates one subtask: The first layer of nonlinear equality describes basic tracking task based on forward kinematics; The second layer and third layer are inequalities, which describe joint angle upper and lower limits; The last layer is a linear equality with respect to joint angle velocity, which could be designed by user to describe other task, such as posture control. To solve this complicated four-layered time-variant problem, it is converted as single-layered equation based on the zeroing neural dynamics method. Then, continuous-time solution is proposed. Furthermore, discrete-time algorithm is proposed based on a third-order time-discretization formula and continuous-time solution. Numerical experiments illustrate the effectiveness and superiority compared to existing work.
{"title":"Robot Manipulator Control via Solving Four-Layered Time-Variant Equations Including Linear, Nonlinear Equalities and Inequalities","authors":"Xinhui Zhu, Li Zhang, Yang Shi, Jing Wang, Jian Li","doi":"10.1109/ICICIP53388.2021.9642176","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642176","url":null,"abstract":"Robot manipulator control is a complicated multi-tasking problem in reality. It includes not only basic tracking task, but also additional tasks, such as conquering joint angle limits, posture control. However, most existing works only consider the goal of tracking and formulate it as single-layered time-variant problems, which leads to impracticality. In this work, robot manipulator control problem is formulated as four-layered time-variant equations including linear, nonlinear equalities and inequalities. Each layer formulates one subtask: The first layer of nonlinear equality describes basic tracking task based on forward kinematics; The second layer and third layer are inequalities, which describe joint angle upper and lower limits; The last layer is a linear equality with respect to joint angle velocity, which could be designed by user to describe other task, such as posture control. To solve this complicated four-layered time-variant problem, it is converted as single-layered equation based on the zeroing neural dynamics method. Then, continuous-time solution is proposed. Furthermore, discrete-time algorithm is proposed based on a third-order time-discretization formula and continuous-time solution. Numerical experiments illustrate the effectiveness and superiority compared to existing work.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129548755","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642164
Luyang Han, Bolin Liao, Yongjun He, Xiao Xiao
Original zeroing neural network (OZNN) can effectively solve the problem of matrix inversion. Generally, the problem of matrix inversion is solved in the noiseless environment. However, noises are common, OZNN can not solve the problem with harmonic noise interference. Therefore, the integrated enhanced zeroing neural network (IEZNN) is proposed to overcome this difficulty. IEZNN can deal with the harmonic noise interference problem when the time change slightly. But in the case of large amplitude or frequency, IEZNN has not strong ability to tolerate the noise and the convergence speed is relatively slow. Therefore, by adding a novel nonlinear activation for IEZNN, which also has the ability to suppress noise, a dual noise-suppressed ZNN (DNSZNN) is proposed to solve this problem. DNSZNN not only has good noise suppression characteristics, but also can converge in the predefined time. Finally, the experimental results demonstrate that the DNSZNN has the best robustness and convergence performance under the same external harmonic noise interference compared with the OZNN and the IEZNN.
{"title":"Dual Noise-Suppressed ZNN with Predefined-Time Convergence and its Application in Matrix Inversion","authors":"Luyang Han, Bolin Liao, Yongjun He, Xiao Xiao","doi":"10.1109/ICICIP53388.2021.9642164","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642164","url":null,"abstract":"Original zeroing neural network (OZNN) can effectively solve the problem of matrix inversion. Generally, the problem of matrix inversion is solved in the noiseless environment. However, noises are common, OZNN can not solve the problem with harmonic noise interference. Therefore, the integrated enhanced zeroing neural network (IEZNN) is proposed to overcome this difficulty. IEZNN can deal with the harmonic noise interference problem when the time change slightly. But in the case of large amplitude or frequency, IEZNN has not strong ability to tolerate the noise and the convergence speed is relatively slow. Therefore, by adding a novel nonlinear activation for IEZNN, which also has the ability to suppress noise, a dual noise-suppressed ZNN (DNSZNN) is proposed to solve this problem. DNSZNN not only has good noise suppression characteristics, but also can converge in the predefined time. Finally, the experimental results demonstrate that the DNSZNN has the best robustness and convergence performance under the same external harmonic noise interference compared with the OZNN and the IEZNN.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"231 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122864073","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642196
Kangning Yin, Rui Zhu, Shaoqi Hou, Guangqiang Yin
Sleep staging has a strong reference value in modern medicine for doctors to judge patients’ physical and mental state and provide treatment advice. However, in reality, according to the original information of sleep Electroencephalogram (EEG), it is difficult for doctors to manually judge, and sleep staging samples are difficult to obtain, so the data is few. At the same time, the robustness of the sleep staging model obtained only by individual learning is poor. In order to solve the problem of using fuzzy few samples to design the sleep staging prediction model to provide accurate sleep staging information for doctors, an unsupervised auxiliary algorithm model is designed. Firstly, according to the data characteristics of sleep EEG signals, low-pass filtering and fast Fourier transform were performed on the EEG signals recorded during sleep. Sleep stages are performed according to the frequency parameters, and normalization is performed to highlight the wave characteristics of different components. Secondly, due to the existence of different sample data in each stage, unsupervised samples are classified and corrected by K-Means clustering method, and a more robust model is trained under the premise of ensuring the diversity of training samples. Finally, the data set divided by clustering is sent to Support Vector Machine (SVM) classification learning, and the Gaussian kernel function is used to achieve high-dimensional mapping, which can reduce the impact of deviation from the center data on the sample center. The sleep staging classification algorithm designed in this paper can classify the sleep staging under the condition of fuzzy few samples, in the case of equal proportion of training set and test set, the correct rate is higher than 90 %, and in very few samples, the classification accuracy is more than 85 %.
{"title":"Unsupervised Assisted Sleep staging Classification Algorithm under Fuzzy Few Samples","authors":"Kangning Yin, Rui Zhu, Shaoqi Hou, Guangqiang Yin","doi":"10.1109/ICICIP53388.2021.9642196","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642196","url":null,"abstract":"Sleep staging has a strong reference value in modern medicine for doctors to judge patients’ physical and mental state and provide treatment advice. However, in reality, according to the original information of sleep Electroencephalogram (EEG), it is difficult for doctors to manually judge, and sleep staging samples are difficult to obtain, so the data is few. At the same time, the robustness of the sleep staging model obtained only by individual learning is poor. In order to solve the problem of using fuzzy few samples to design the sleep staging prediction model to provide accurate sleep staging information for doctors, an unsupervised auxiliary algorithm model is designed. Firstly, according to the data characteristics of sleep EEG signals, low-pass filtering and fast Fourier transform were performed on the EEG signals recorded during sleep. Sleep stages are performed according to the frequency parameters, and normalization is performed to highlight the wave characteristics of different components. Secondly, due to the existence of different sample data in each stage, unsupervised samples are classified and corrected by K-Means clustering method, and a more robust model is trained under the premise of ensuring the diversity of training samples. Finally, the data set divided by clustering is sent to Support Vector Machine (SVM) classification learning, and the Gaussian kernel function is used to achieve high-dimensional mapping, which can reduce the impact of deviation from the center data on the sample center. The sleep staging classification algorithm designed in this paper can classify the sleep staging under the condition of fuzzy few samples, in the case of equal proportion of training set and test set, the correct rate is higher than 90 %, and in very few samples, the classification accuracy is more than 85 %.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130049238","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642206
Tiange Ye, Rushi Lan, Xiaonan Luo
In this paper, we propose a new method for colon cancer classification from histopathological images, which can automatically analyze a given whole slide image (WSI). We usually classify cancer classification by referring a WSI, which is typically 20000 × 20000 pixels. The cost of obtaining WSIs with annotating cancer regions is very high. Multiple-instance learning (MIL) is a variant of supervised learning in which the instances in a bag share a single class label. That is, MIL only needs unannotated WSI. In recent years, MIL has developed a hard attention mechanism which has achieved good performance. However, this hard attention mechanism cannot notice the interior of each patch, i.e., it lacks soft attention mechanism. In this paper, a soft, sequential, spatial, top-down attention mechanism (which we abbreviate as S3TA) is used to make up for the lack of MIL attention mechanism. Finally, our experiments show that by varying the number of attention steps in S3TA, we achieved a better accuracy of 93.6% than the old model.
{"title":"Multiple-instance CNN Improved by S3TA for Colon Cancer Classification with Unannotated Histopathological Images","authors":"Tiange Ye, Rushi Lan, Xiaonan Luo","doi":"10.1109/ICICIP53388.2021.9642206","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642206","url":null,"abstract":"In this paper, we propose a new method for colon cancer classification from histopathological images, which can automatically analyze a given whole slide image (WSI). We usually classify cancer classification by referring a WSI, which is typically 20000 × 20000 pixels. The cost of obtaining WSIs with annotating cancer regions is very high. Multiple-instance learning (MIL) is a variant of supervised learning in which the instances in a bag share a single class label. That is, MIL only needs unannotated WSI. In recent years, MIL has developed a hard attention mechanism which has achieved good performance. However, this hard attention mechanism cannot notice the interior of each patch, i.e., it lacks soft attention mechanism. In this paper, a soft, sequential, spatial, top-down attention mechanism (which we abbreviate as S3TA) is used to make up for the lack of MIL attention mechanism. Finally, our experiments show that by varying the number of attention steps in S3TA, we achieved a better accuracy of 93.6% than the old model.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116691800","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642193
Yu Chen, Liang Chen, Yan Wang, Yu Zheng, Huade Su
To reduce the missed inspection rate of unqualified welded seams of the hull, a model based on EasyEnsemble and XGBoost algorithm is proposed to predict the ultrasonic inspection results of welds. Based on historical data of weld ultrasonic inspection, parameters related to the welding quality were selected and these parameters were processed by feature engineering such as normalization and coding. Then effective features were extracted as the model input by principal component analysis (PCA). Considering the low recall of negative samples caused by extremely unbalanced sample data distribution, the EasyEnsemble algorithm was adopted to obtain a balanced training sample set and XGBoost algorithm was used as the base classification model of EasyEnsemble algorithm. The validity of the proposed model was proved by the experiment, the recall of negative samples was greatly improved and the missed inspection rate of unqualified welds was reduced.
{"title":"Application Research on Prediction of Weld Ultrasonic Inspection Results Based on EasyEnsemble and XGBoost Algorithm","authors":"Yu Chen, Liang Chen, Yan Wang, Yu Zheng, Huade Su","doi":"10.1109/ICICIP53388.2021.9642193","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642193","url":null,"abstract":"To reduce the missed inspection rate of unqualified welded seams of the hull, a model based on EasyEnsemble and XGBoost algorithm is proposed to predict the ultrasonic inspection results of welds. Based on historical data of weld ultrasonic inspection, parameters related to the welding quality were selected and these parameters were processed by feature engineering such as normalization and coding. Then effective features were extracted as the model input by principal component analysis (PCA). Considering the low recall of negative samples caused by extremely unbalanced sample data distribution, the EasyEnsemble algorithm was adopted to obtain a balanced training sample set and XGBoost algorithm was used as the base classification model of EasyEnsemble algorithm. The validity of the proposed model was proved by the experiment, the recall of negative samples was greatly improved and the missed inspection rate of unqualified welds was reduced.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116034301","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642169
S. Cong, Kezhi Li
The paper synthesizes the local neural networks. Network structures and their activation functions of three local networks CMAC, B-spline, RBF that are often used to approach functions are analyzed and compared in detail. The network structure of ART-2 is also discussed. Based on the fuzzy system of these local networks, the paper depicts their fuzzy structures and performances. The study and analysis in the paper are useful to instruct to select and design the local neural networks.
{"title":"Analysis and Comparison of the Structure and Performance of Local Neural Networks","authors":"S. Cong, Kezhi Li","doi":"10.1109/ICICIP53388.2021.9642169","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642169","url":null,"abstract":"The paper synthesizes the local neural networks. Network structures and their activation functions of three local networks CMAC, B-spline, RBF that are often used to approach functions are analyzed and compared in detail. The network structure of ART-2 is also discussed. Based on the fuzzy system of these local networks, the paper depicts their fuzzy structures and performances. The study and analysis in the paper are useful to instruct to select and design the local neural networks.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128105237","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642162
Meng Zhang, Yao Xiao, Xiaoling Song, Xiangguang Dai, Nian Zhang
There are balanced priorities in various engineering fields (e.g. medicine, statistics, artificial intelligence, and economics, etc.). Some clustering algorithms cannot maintain the natural balanced structure of data. This paper proposes a soft-balanced clustering framework, which can achieve a balanced clustering for each cluster. The model can be formulated d as a mixed-integer optimization problem. We transform the problem into several subproblems and utilize PSO to search the global solution. Experiments show that the proposed algorithm can achieve satisfactory clustering results than other clustering algorithms.
{"title":"Fast Particle Swarm optimization for Balanced Clustering","authors":"Meng Zhang, Yao Xiao, Xiaoling Song, Xiangguang Dai, Nian Zhang","doi":"10.1109/ICICIP53388.2021.9642162","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642162","url":null,"abstract":"There are balanced priorities in various engineering fields (e.g. medicine, statistics, artificial intelligence, and economics, etc.). Some clustering algorithms cannot maintain the natural balanced structure of data. This paper proposes a soft-balanced clustering framework, which can achieve a balanced clustering for each cluster. The model can be formulated d as a mixed-integer optimization problem. We transform the problem into several subproblems and utilize PSO to search the global solution. Experiments show that the proposed algorithm can achieve satisfactory clustering results than other clustering algorithms.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124908309","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 : 2021-12-03DOI: 10.1109/ICICIP53388.2021.9642200
Shengjie Chen, Mao Ye
With the great success of deep learning network, compressed video quality enhancement methods based on deep learning are mushrooming. Most of these methods ignore the correlation between frames and do not make full use of the information of adjacent frames. We propose a two-stage multi-frame cooperative quality enhancement network. Our method consist of two main modules: motion compensation network and quality enhancement network. We use a two-stage enhanced structure to make full use of high-quality frames information and realize the multi-frame cooperative enhancement of a Group of Pictures(GOP), fully considering the correlation between frames. The experimental results on the HEVC standard test sequences show that the proposed method is improved by about 10% compared with MFQE2.0.
{"title":"Two-stage Multi-frame Cooperative Quality Enhancement on Compressed Video","authors":"Shengjie Chen, Mao Ye","doi":"10.1109/ICICIP53388.2021.9642200","DOIUrl":"https://doi.org/10.1109/ICICIP53388.2021.9642200","url":null,"abstract":"With the great success of deep learning network, compressed video quality enhancement methods based on deep learning are mushrooming. Most of these methods ignore the correlation between frames and do not make full use of the information of adjacent frames. We propose a two-stage multi-frame cooperative quality enhancement network. Our method consist of two main modules: motion compensation network and quality enhancement network. We use a two-stage enhanced structure to make full use of high-quality frames information and realize the multi-frame cooperative enhancement of a Group of Pictures(GOP), fully considering the correlation between frames. The experimental results on the HEVC standard test sequences show that the proposed method is improved by about 10% compared with MFQE2.0.","PeriodicalId":435799,"journal":{"name":"2021 11th International Conference on Intelligent Control and Information Processing (ICICIP)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124205498","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：