模式识别与人工智能
Home      About Journal      Editorial Board      Instructions      Ethics Statement      Contact Us                   中文
Pattern Recognition and Artificial Intelligence
22 Judgement and Disposal of Academic Misconduct Article
22 Copyright Transfer Agreement
22 Proof of Confidentiality
22 Requirements for Electronic Version
More....
22 Chinese Association of Automation
22 National ResearchCenter for Intelligent Computing System
22 Institute of Intelligent Machines,Chinese Academy of Sciences
More....
 
 
2020 Vol.33 Issue.7, Published 2020-07-25
Papers and Reports    Researches and Applications    Surveys and Reviews   
   
Papers and Reports
575 Interactive Meticulous Flower Coloring Algorithm via Attention Guidance
LI Yuan, CHEN Zhaojiong, YE Dongyi
The process of manually rendering traditional Chinese meticulous flower painting is complicated and highly skilled. The existing automatic line drawing colorization is difficult to generate natural and reasonable gradient effect. On the basis of condition generative adversarial network(CGAN), an interactive meticulous flower coloring algorithm via attention guidance is proposed to accomplish the colorization of meticulous flowers from line drawing. A color attention map depicting the color category and layout of flowers is designed to guide the proposed network to learn important color features in the training stage. The color attention map is considered as the means of interaction between the user and the system for color design in the application stage. In the network structure design, a local color-coding sub-network is constructed and trained to encode the flower color attention map. The encoded information is introduced into the conditional normalization process of each layer of the generator as an affine parameter to accomplish learning and controlling of colors. Since the depth features emphasize global semantic information, the local high-frequency information reflecting line features might be lost. A cross-layer connection structure is introduced into the generator network to strengthen the learning of line features. Experimental results show that the proposed algorithm renders line drawing of flowers better into meticulous flowers and the generated images are accordant with the color distribution and characteristics of real meticulous flowers with good artistic reality and appreciation.
2020 Vol. 33 (7): 575-587 [Abstract] ( 704 ) [HTML 1KB] [ PDF 5146KB] ( 497 )
588 Surface Electromyography Classification Method Based on Temporal Two-Dimensionalization and Convolution Feature Fusion
LUO Junjin, WANG Wanliang, WANG Zheng, LIU Honghai
The traditional pattern recognition methods are prone to ignore characteristics of non-linearity and timing in the classification of surface electromyography(sEMG). Aiming at this problem, a sEMG signal classification method based on temporal two-dimensionalization and convolution feature fusion is proposed. Temporal two-dimensionalization is realized by Gramian angular field conversion to preserve the time dependence and correlation of original time series of sEMG. To highlight the local information and fully retain details simultaneously, a capsule network and a convolutional neural network are introduced to extract features together. In addition, the feature fusion is performed to realize the gesture recognition under different conditions. Experimental results show that the proposed method is more robust than other classification methods and it effectively enhances the electrode offset and the overall recognition level of hand movements facing new objects.
2020 Vol. 33 (7): 588-599 [Abstract] ( 861 ) [HTML 1KB] [ PDF 1186KB] ( 1545 )
600 Three-Way Recommendation Based on Trust Transfer Mechanism
QIN Qin, ZHANG Hengru
To solve the sparseness problem of direct trust relationships in social networks and reduce the recommendation cost of traditional collaborative filtering algorithms, a three-way recommendation algorithm based on trust transfer mechanism is proposed. Firstly, a trust transfer mechanism is built to obtain the user's indirect trust relationships to expand the user's social networks. Secondly, the bipartite graph network structure is applied to calculate the bidirectional influence factors between users. Then, the bidirectional influence factors are regarded as the constraint term to design a new objective function to participate in the matrix factorization. Finally, the misclassification cost and promotion cost are taken into account in the recommendation process by introducing three-way decision, and thus a three-way recommendation algorithm based on objective function is presented. Experimental results on Filmtrust and Epinions datasets indicate that the proposed algorithm is superior to the traditional collaborative filtering algorithms.
2020 Vol. 33 (7): 600-609 [Abstract] ( 553 ) [HTML 1KB] [ PDF 594KB] ( 343 )
610 Hesitant Fuzzy Multi-attribute Decision Making Based on Conflict Analysis
ZHANG Huimin, LI Xiaonan
Defining distance measure and calculating attribute weight are two key points of multi-attribute decision making based on hesitant fuzzy distance. During the process of defining distance measure of two hesitant fuzzy numbers, the original meaning of the shorter hesitant fuzzy number is changed when some same elements are added to it. In this paper, two hesitant fuzzy numbers are extended to the same length simultaneously. Then, the conflict analysis model of hesitant fuzzy information system is established according to the conflict analysis theory by Pawlak, and the method for calculating attribute weights based on the degree of conflict is proposed. Finally, the specific method of multi-attribute decision-making under hesitant fuzzy information systems is presented. The effectiveness and the feasibility of the proposed method are exemplified on the basis of a case study on a enterprise development plan.
2020 Vol. 33 (7): 610-618 [Abstract] ( 469 ) [HTML 1KB] [ PDF 519KB] ( 334 )
Surveys and Reviews
619 Concept Lattice Theory and Method and Their Research Prospect
LI Jinhai, WEI Ling, ZHANG Zhuo, ZHAI Yanhui, ZHANG Tao, ZHI Huilai, MI Yunlong

Concept lattice theory and method are the basic topics in the study of formal concept analysis, and important achievements are obtained. The previous study mainly focus on the generalization of concept lattice models, concept lattice construction, concept lattice reduction, concept lattice based rule acquisition, conceptual knowledge space, granular computing method for concept lattice and concept lattice applications. To further promote the study and the development of formal concept analysis theory, the existing research on theories and methods of concept lattice is summarized in detail, and an outlook is also produced for the researchers. Especially, it is pointed out that there are some key scientific problems in the above researches, a theoretical analysis of these problems is presented, and some preliminary research thoughts are provided as well. The obtained results offer a useful piece of advice for solving the problems in the future.
2020 Vol. 33 (7): 619-642 [Abstract] ( 1285 ) [HTML 1KB] [ PDF 1362KB] ( 750 )
Researches and Applications
643 Semi-supervised Nuclei Segmentation Based on Consistency Regularization Constraint
SHU Jianhua, NIAN Fudong, LÜ Gang
Aiming at the high cost of medical image data acquisition with high quality annotation, a semi-supervised nuclei segmentation algorithm based on consistency regularization constraint is proposed. Firstly, two master and slave networks with the same structure are constructed, and the same random initialization parameters are assigned to them. Then, the labeled and unlabeled training data are randomly selected to input into the master and slave networks. Regularization term is utilized to constrain the training of master and slave networks to keep the output results consistent. The parameters of master network are optimized by gradient descent method, and the parameters of the slave network are optimized by the exponential moving average of the parameters of master network in each iteration batches. Experiments on public datasets verify the effectiveness of the proposed algorithm.
2020 Vol. 33 (7): 643-652 [Abstract] ( 553 ) [HTML 1KB] [ PDF 836KB] ( 421 )
653 Chinese Named Entity Recognition Method Based on Machine Reading Comprehension
LIU Yiyang, YU Zhengtao, GAO Shengxiang, GUO Junjun, ZHANG Yafei, NIE Bingge
The existing named entity recognition methods mainly consider the context information in a single sentence, rather than the impact of document-level context. Aiming at this problem, a Chinese named entity recognition method based on reading comprehension is proposed, and the idea of reading comprehension is utilized to fully mine document-level context features to support entity recognition. Firstly, for each type of entity, the entity recognition task is transformed into a question and answer task, and a triple of question, text and entity answer is constructed. Then, the triple information is passed through BERT pre-training and convolutional neural network to capture document-level text context information. Finally, the entity answer prediction is realized through the binary classifier. The experiment of named entity recognition on MSRA dataset, People's Daily public dataset and self-built dataset shows the better performance of the proposed method and the better effect of reading comprehension on entity recognition.
2020 Vol. 33 (7): 653-659 [Abstract] ( 916 ) [HTML 1KB] [ PDF 770KB] ( 606 )
660 Butterfly Optimization Algorithm Combining Sine Cosine and Iterative Chaotic Map with Infinite Collapses
WANG Yirou, ZHANG Damin

To solve the problem of low precision and convergence speed of butterfly optimization algorithm, a butterfly optimization algorithm combining sine cosine and iterative chaotic map with infinite collapses(SIBOA) is proposed. Firstly, iterative chaotic map with infinite collapses is utilized to initialize the individual state of butterflies to avoid the algorithm falling into local optimum. Secondly, sine and cosine operators are introduced into the cognitive flight part to balance local and global search capability of the algorithm. Finally, the power index dependent on the size of fragrance is improved to adjust its absorption degree to obtain a better optimal solution. The results on 8 benchmark functions show that SIBOA achieves better global search capability, solution robustness and optimization accuracy with a higher convergence speed.
2020 Vol. 33 (7): 660-669 [Abstract] ( 547 ) [HTML 1KB] [ PDF 766KB] ( 904 )
模式识别与人工智能
 

Supervised by
China Association for Science and Technology
Sponsored by
Chinese Association of Automation
NationalResearchCenter for Intelligent Computing System
Institute of Intelligent Machines, Chinese Academy of Sciences
Published by
Science Press
 
Copyright © 2010 Editorial Office of Pattern Recognition and Artificial Intelligence
Address: No.350 Shushanhu Road, Hefei, Anhui Province, P.R. China Tel: 0551-65591176 Fax:0551-65591176 Email: bjb@iim.ac.cn
Supported by Beijing Magtech 　Email:support@magtech.com.cn