Pub Date : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302201
Ran Fang, Yinan Lu, Xiaoni Liu, Zhuo Liu
FCM algorithm is a popular algorithm for medical image segmentation. The precise process of segmenting brain tissue images becomes more challenging in the presence of noise and other image artifacts. An improved adaptively regularized kernel FCM method is proposed in this paper. The spatial constraint function of membership is introduced to enhance clustering by adjusting the degree of influence between pixels and clustering centers. Experimental results on the brain images with different types and levels of noises demonstrate that the improved algorithm increases the accuracy of segmentation compared with the other soft clustering algorithms.
{"title":"Segmentation of brain MR images using an adaptively regularized kernel FCM algorithm with spatial constraints","authors":"Ran Fang, Yinan Lu, Xiaoni Liu, Zhuo Liu","doi":"10.1109/CISP-BMEI.2017.8302201","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302201","url":null,"abstract":"FCM algorithm is a popular algorithm for medical image segmentation. The precise process of segmenting brain tissue images becomes more challenging in the presence of noise and other image artifacts. An improved adaptively regularized kernel FCM method is proposed in this paper. The spatial constraint function of membership is introduced to enhance clustering by adjusting the degree of influence between pixels and clustering centers. Experimental results on the brain images with different types and levels of noises demonstrate that the improved algorithm increases the accuracy of segmentation compared with the other soft clustering algorithms.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"114 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77633223","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8301960
Ye Yuan, Liangzhuo Xie, Yewen Zhu, Sheng Wang, Zhemin Zhuang
A new SAR image de-noising approach that uses the local properties analysis of SAR image and the discrete nonseparable shearlet transform (DNST) is proposed in this paper. According to the local properties analysis method, the SAR image is divided into homogeneous region, non-homogeneous region and target region. The homogeneous region uses the average filter to de-noising. The non-homogeneous region uses the DNST transform to de-noising and target region is reserved directly. The experimental results show that the proposed approach can efficiently reduce the speckle noises and improve the edge-preserving ability.
{"title":"SAR image de-noising using local properties analysis and discrete non-separable shearlet transform","authors":"Ye Yuan, Liangzhuo Xie, Yewen Zhu, Sheng Wang, Zhemin Zhuang","doi":"10.1109/CISP-BMEI.2017.8301960","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8301960","url":null,"abstract":"A new SAR image de-noising approach that uses the local properties analysis of SAR image and the discrete nonseparable shearlet transform (DNST) is proposed in this paper. According to the local properties analysis method, the SAR image is divided into homogeneous region, non-homogeneous region and target region. The homogeneous region uses the average filter to de-noising. The non-homogeneous region uses the DNST transform to de-noising and target region is reserved directly. The experimental results show that the proposed approach can efficiently reduce the speckle noises and improve the edge-preserving ability.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"33 7-8 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77660004","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302319
S. Chang, Anne Yenching Liu, Yu-Teng Jang
We investigated the employee's trust, commitment and compliance in the practice of information monitoring — an important information security management (ISM) issue for healthcare organizations to prevent employee misuse behaviors and safeguard sensitive information (e.g. medical records). Many studies have explored similar issues of information monitoring practice, but unfortunately they mostly overlook the domain attributes of ISM in reality. We studied the theories of privacy, trust, commitment and compliance to formulate a model for explaining the phenomenon observed in real work environment with organizational information monitoring in practice. Our research accomplished an advanced exploration of information monitoring for enhancing organizational ISM practices (ISMP).
{"title":"Exploring trust and information monitoring for information security management","authors":"S. Chang, Anne Yenching Liu, Yu-Teng Jang","doi":"10.1109/CISP-BMEI.2017.8302319","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302319","url":null,"abstract":"We investigated the employee's trust, commitment and compliance in the practice of information monitoring — an important information security management (ISM) issue for healthcare organizations to prevent employee misuse behaviors and safeguard sensitive information (e.g. medical records). Many studies have explored similar issues of information monitoring practice, but unfortunately they mostly overlook the domain attributes of ISM in reality. We studied the theories of privacy, trust, commitment and compliance to formulate a model for explaining the phenomenon observed in real work environment with organizational information monitoring in practice. Our research accomplished an advanced exploration of information monitoring for enhancing organizational ISM practices (ISMP).","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"24 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78170786","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302163
Wang Yunpeng, Wang Wei, L. Jianhua, Liu Yongheng, Zhao Yijie, Li Xing, Guan-jun Lei, Liu Guozhong
Farmers are prone to be scratched when they are picking Chinese peppers by hands because of the thorns. Therefore, the safe and effective picking tools are extremely significant for the workers. But there is no one Chinese pepper picking tool that is really suitable for farmers. In this paper, four Sichuan pepper picking tool schemes are designed based on the TRIZ theory and they are evaluated by the triangular fuzzy number complementary judgment matrix. The evaluation indexes are marked by four raters. Calculating utility value and then we select the best scheme. And this paper may provide an evaluation method for the other product Designs.
{"title":"Chinese pepper picking tool designs and evaluations based on the TRIZ theory and the triangular fuzzy number","authors":"Wang Yunpeng, Wang Wei, L. Jianhua, Liu Yongheng, Zhao Yijie, Li Xing, Guan-jun Lei, Liu Guozhong","doi":"10.1109/CISP-BMEI.2017.8302163","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302163","url":null,"abstract":"Farmers are prone to be scratched when they are picking Chinese peppers by hands because of the thorns. Therefore, the safe and effective picking tools are extremely significant for the workers. But there is no one Chinese pepper picking tool that is really suitable for farmers. In this paper, four Sichuan pepper picking tool schemes are designed based on the TRIZ theory and they are evaluated by the triangular fuzzy number complementary judgment matrix. The evaluation indexes are marked by four raters. Calculating utility value and then we select the best scheme. And this paper may provide an evaluation method for the other product Designs.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"114 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79448530","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302049
Yuzhong Liu, Qiyang Zhao, Cheng Jiang
Generative Adversarial Net is a frontier method of generative models for images, audios and videos. In this paper, we focus on conditional image generation and introduce conditional Feature-Matching Generative Adversarial Net to generate images from category labels. By visualizing state-of-art discriminative conditional generative models, we find these networks do not gain clear semantic concepts. Thus we design the loss function in the light of metric learning to measure semantic distance. The proposed model is evaluated on several well-known datasets. It is shown to be of higher perceptual quality and better diversity then existing generative models.
{"title":"Conditional image generation using feature-matching GAN","authors":"Yuzhong Liu, Qiyang Zhao, Cheng Jiang","doi":"10.1109/CISP-BMEI.2017.8302049","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302049","url":null,"abstract":"Generative Adversarial Net is a frontier method of generative models for images, audios and videos. In this paper, we focus on conditional image generation and introduce conditional Feature-Matching Generative Adversarial Net to generate images from category labels. By visualizing state-of-art discriminative conditional generative models, we find these networks do not gain clear semantic concepts. Thus we design the loss function in the light of metric learning to measure semantic distance. The proposed model is evaluated on several well-known datasets. It is shown to be of higher perceptual quality and better diversity then existing generative models.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"16 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81725682","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302328
Yong Liu
In order to prevent the individual neural networks from becoming similar in the long learning period of negative correlation learning for designing neural network ensembles, two approaches were adopted in this paper. The first approach is to replace large neural networks with small neural networks in neural network ensembles. Samll neural networks would be more practical in the real applications when the capability is limited. The second approach is to introduce random separation learning in negative correlation learning for each small neural network. The idea of random separation learning is to let each individual neural network learn differently on the randomly separated subsets of the given training samples. It has been found that the small neural networks could easily become weak and different each other by negative correlation learning with random separation learning. After applying large number of small neural networks for neural network ensembles, two combination methods were used to generate the output of the neural network ensembles while their performance had been compared.
{"title":"Random separation learning for neural network ensembles","authors":"Yong Liu","doi":"10.1109/CISP-BMEI.2017.8302328","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302328","url":null,"abstract":"In order to prevent the individual neural networks from becoming similar in the long learning period of negative correlation learning for designing neural network ensembles, two approaches were adopted in this paper. The first approach is to replace large neural networks with small neural networks in neural network ensembles. Samll neural networks would be more practical in the real applications when the capability is limited. The second approach is to introduce random separation learning in negative correlation learning for each small neural network. The idea of random separation learning is to let each individual neural network learn differently on the randomly separated subsets of the given training samples. It has been found that the small neural networks could easily become weak and different each other by negative correlation learning with random separation learning. After applying large number of small neural networks for neural network ensembles, two combination methods were used to generate the output of the neural network ensembles while their performance had been compared.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"108 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85217703","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8301964
Kai Cui, Hua Cai, Yao Zhang, Huan Chen
Nowadays, face recognition research has been widely concerned, and facial face feature point positioning, that is, face alignment is an important part of the face recognition process, the accuracy of positioning and positioning speed directly affect the face recognition effect. The face alignment task in the real scene becomes a very difficult problem because of the presence of factors such as different pose, expression, illumination and partial occlusion in face images. Aiming at these problems, this paper presents a face alignment method based on SURF of Scale Invariant Feature Transform, which can quickly detect and characterize the key points of face image. In addition, We use a coarse to fine auto-encoder networks to implement complex non-linear mapping of face to face shape. Finally, By comparing the AFLW data set, It shows that the mean error rate of this method is 1.84%-2.74% lower than that of the traditional method, and It also has a good effect in the calculation speed.
{"title":"A face alignment method based on SURF features","authors":"Kai Cui, Hua Cai, Yao Zhang, Huan Chen","doi":"10.1109/CISP-BMEI.2017.8301964","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8301964","url":null,"abstract":"Nowadays, face recognition research has been widely concerned, and facial face feature point positioning, that is, face alignment is an important part of the face recognition process, the accuracy of positioning and positioning speed directly affect the face recognition effect. The face alignment task in the real scene becomes a very difficult problem because of the presence of factors such as different pose, expression, illumination and partial occlusion in face images. Aiming at these problems, this paper presents a face alignment method based on SURF of Scale Invariant Feature Transform, which can quickly detect and characterize the key points of face image. In addition, We use a coarse to fine auto-encoder networks to implement complex non-linear mapping of face to face shape. Finally, By comparing the AFLW data set, It shows that the mean error rate of this method is 1.84%-2.74% lower than that of the traditional method, and It also has a good effect in the calculation speed.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"115 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77083843","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302147
Guang Li, Maolin Li, Dan Liu, Guanghua Xu, Shi-Lin Zhou
The decision process of the traditional mechanical equipment fault diagnosis method is invisible; it is similar to the black box operation and difficult to find the hidden knowledge in the data. Aiming at this problem, a fault diagnosis method of mechanical equipment is proposed based on data visualization. Firstly the data is flattened based on the constellation, and taking into account the different contribution that each data plays, the weight of each feature data is optimized by genetic algorithm, and then the fault diagnosis model based on data visualization is constructed by using the boundary form of the plane point set. Finally the results of the experiments on gearbox test experiment reveal that the proposed method is superior to the K-Neighborhood method and accurate for the fault diagnosis.
{"title":"Fault diagnosis of mechanical equipment based on data visualization","authors":"Guang Li, Maolin Li, Dan Liu, Guanghua Xu, Shi-Lin Zhou","doi":"10.1109/CISP-BMEI.2017.8302147","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302147","url":null,"abstract":"The decision process of the traditional mechanical equipment fault diagnosis method is invisible; it is similar to the black box operation and difficult to find the hidden knowledge in the data. Aiming at this problem, a fault diagnosis method of mechanical equipment is proposed based on data visualization. Firstly the data is flattened based on the constellation, and taking into account the different contribution that each data plays, the weight of each feature data is optimized by genetic algorithm, and then the fault diagnosis model based on data visualization is constructed by using the boundary form of the plane point set. Finally the results of the experiments on gearbox test experiment reveal that the proposed method is superior to the K-Neighborhood method and accurate for the fault diagnosis.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"18 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80933614","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302087
Bihan Zhang, Chuchu Ding, Wei Yan, Li Guo, Jun Wang, F. Hou
In this paper, we symbolize two kinds of different channels of Magnetoencephalography(MEG) and analyze their coupling relationship using symbolic transfer entropy algorithm. We record MEG signals from six depressive disorders and nine healthy subjects stimulated by positive, neutral, and negative emotional pictures and explore coupling relationship of different MEG channels. The results show that there are obvious differences on correlations between two channels of MLP32 and MRP32 with positive emotional stimulus, MLP31 and MRP31 with neutral emotional stimulus, MLP53 and MRP53 with negative emotional stimulus. In general, these channels have more correlation in patients with major depression, and can be able to distinguish depression patient from crowd. It also shows that the research of symbolic transfer entropy in MEG channel can distinguish the difference between normal and case samples, which of significance for clinical pathological estimation and diagnosis.
{"title":"Analysis of Magnetoencephalography based on symbolic transfer entropy","authors":"Bihan Zhang, Chuchu Ding, Wei Yan, Li Guo, Jun Wang, F. Hou","doi":"10.1109/CISP-BMEI.2017.8302087","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302087","url":null,"abstract":"In this paper, we symbolize two kinds of different channels of Magnetoencephalography(MEG) and analyze their coupling relationship using symbolic transfer entropy algorithm. We record MEG signals from six depressive disorders and nine healthy subjects stimulated by positive, neutral, and negative emotional pictures and explore coupling relationship of different MEG channels. The results show that there are obvious differences on correlations between two channels of MLP32 and MRP32 with positive emotional stimulus, MLP31 and MRP31 with neutral emotional stimulus, MLP53 and MRP53 with negative emotional stimulus. In general, these channels have more correlation in patients with major depression, and can be able to distinguish depression patient from crowd. It also shows that the research of symbolic transfer entropy in MEG channel can distinguish the difference between normal and case samples, which of significance for clinical pathological estimation and diagnosis.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"36 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80938058","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 : 2017-10-01DOI: 10.1109/CISP-BMEI.2017.8302170
Guangwen Liu, Hua Cai, Yang Yang, Z. Geng
Aiming at the problems of color distortion and halos existing in video images in the process of transplanting dark channel prior algorithm of de-hazing into FPGA and considering the algorithm implementation in FPGA and the real-time requirement of video processing, an improvement algorithm base on FPGA is proposed in this paper. By refining the dark channel to obtain the final transmittance for eliminating the halos and by effectively segmenting the bright regions like the sky regions, the color distortion can be eliminated through the non-linear compensation for the regional transmittance. Through the verification of MATLAB simulation and the experiment of FPGA hardware, the system can effectively solve above problems. The experimental results show that on the premise of ensuring the output video quality, the system can better de-haze and eliminate halos in video images.
{"title":"Improvement and implementation of video image de-hazing algorithm based on FPGA","authors":"Guangwen Liu, Hua Cai, Yang Yang, Z. Geng","doi":"10.1109/CISP-BMEI.2017.8302170","DOIUrl":"https://doi.org/10.1109/CISP-BMEI.2017.8302170","url":null,"abstract":"Aiming at the problems of color distortion and halos existing in video images in the process of transplanting dark channel prior algorithm of de-hazing into FPGA and considering the algorithm implementation in FPGA and the real-time requirement of video processing, an improvement algorithm base on FPGA is proposed in this paper. By refining the dark channel to obtain the final transmittance for eliminating the halos and by effectively segmenting the bright regions like the sky regions, the color distortion can be eliminated through the non-linear compensation for the regional transmittance. Through the verification of MATLAB simulation and the experiment of FPGA hardware, the system can effectively solve above problems. The experimental results show that on the premise of ensuring the output video quality, the system can better de-haze and eliminate halos in video images.","PeriodicalId":6474,"journal":{"name":"2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"10 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2017-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78561630","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：