Survey and classification of spatial object manipulation techniques

标准评析《基于国际功能、残疾和健康分类评价量表应用技术指南》团体标准解读■ 徐悦莹1，2 郑洁皎1，2* 曲 冰1，2 沈利岩1，2 李 勇1，2（1.复旦大学附属华东医院康复医学科；2.上海市康复治疗质量控制中心）摘 要：《基于国际功能、残疾和健康分类评价量表应用技术指南》系列团体标准于2020年11月26日正式发布。

该系列团体标准基于ICF理念制定活动和参与评价量表应用技术指南，从“理解交流、身体活动、自我照护、与人相处、生活活动、社会参与”六个领域评价个人的整体健康状况，并在此基础上增加具专病特征的“身体结构”和“身体功能” 两个领域，新增神经、骨骼肌肉、心肺和脏器四个亚专科系统疾病的ICF康复系列评价量表应用技术指南的团体标准。

关键词：国际功能残疾和健康分类，活动和参与，康复结局评价，标准解读DOI编码：10.3969/j.issn.1002-5944.2021.12.019Interpretation of the community standards “Technical Guidelines for Applications of the assessment scale based on International Classificationof Functioning, Disability and Health”XU Yue-ying1,2 ZHENG Jie-jiao1,2* QU Bing1,2 SHEN Li-yan1,2 LI Yong1,2(1. Department of Rehabilitation Medicine, Huadong Hospital Affiliated to Fudan University;2. Shanghai Rehabilitation Medical Quality Control Center)Abstract: The series of association standards including “Technical guidelines for applications of the assessment scale based on International Classification of Functioning, Disability and Health” developed by Huadong Hospital affiliated to Fudan University, Research Institute of Rehabilitation Information and China Rehabilitation Research Center/ WHO-FIC Collaboration Center in China, was published on November 26, 2020. The series of association standards provide technical guidelines for the application of “ICF—Activities and Participation Assessment Scale”, which evaluated the overall health status from six dimensions: “cognition and communication, getting around, self-care, getting along with people, life activities and social participation”. Besides, it added two new dimensions: “body structure” and “body function” with disease characteristics, providing technical guidelines for the application of ICF rehabilitation series assessment scale for four subspecialty diseases in nervous system, musculoskeletal system, cardiopulmonary system, and multiple organ system. This paper aims to make relevant personnel accurately understand assessment contents of the community standards.Keywords: International Classification of Functioning Disability and Health, activities and participation, rehabilitation outcome measure, standard interpretation基金项目：本文是国家重点研发计划“主动健康和老龄化应对”重点专项“基于区块链的老年主动健康智能照护平台研究与应用示范”项目（项目编号：2020YFC2008700）、华东医院院级重点学科建设项目（项目编号：H1160）的研究成果。

The main content of the Action Plan is as follows:1. Strengthen the promotion of standardization work by overall planning. Supporting policies for putting the Outline in place will be improved, and standardization will be incorporated into policies and plans on industry, region, science & technology, trade, etc. The plan of establishing the national standards system will be implemented to facilitate the high-quality development during 2021-2025. Local governments and relevant departments can formulate specific standardization development plans to boost economic and social development.2. Enhance the effective interaction between standardization and technological innovation. The linkage mechanism between major science & technology projects and standardization will be established, and the rate of standards research results achieved by the projects of common key technologies and applied science & technology plans will be counted. Standards research in the key technology fields will be strengthened in the layout of national science & technology plans, and standard will serve as a key indicator in the implementation system of science & technology plans. The projects focusing on standards research are encouraged to carry out the pilot program for preliminary review of national standards proposals. The system of standard essential patents (SEPs) will be completed, and the working mechanism of standards and intellectual property will be established.3. Raise the standards level in equipment manufacturing industry. The standards system for industrial transformation and upgrading will be improved in areas such as integration of informatization and industrialization, and intelligent manufacturing. High-end equipment manufacturing standardization projects will be carried out in key areas such as CNC machine tools, construction machinery, marine equipment, and agricultural machinery to develop and apply standards integrating high-end equipment with information technology, green and low carbon, and modern service. Standardization pilot projects will be conducted in the fields of intelligent, green and service-oriented manufacturing, to create a typical model of promoting high-end development of equipment manufacturing by standards.4. Improve the standards support of modern service industry. Standards in key areas such as smart logistics, cross-border e-commerce and overseas warehouses will be developed to improve the modern circulation system and the multimodal transport standards system. More efforts will be put into the development of national standards for digital transformation of finance, financial risk prevention and control, and financial consumer protection, to accelerate the construction of the green finance standards system. Standards for safety production and packaging management in express delivery sector will be improved, and the pilot and demonstration management system for standardization of service industry will be further enhanced.5. Improve consumer product standards continuously. The plan of improving consumer product standards will be formulated and implemented to expand domestic demands via standards. The development and revision of consumer product safety standards will be strengthened, and more standards for products of children and seniors will be developed. The food quality standards system will be completed. The development of consumer product quality grading standards will be promoted, and the research on consumer experience standards will be strengthened to accelerate the transformation from production-oriented consumer product standards to consumption-oriented ones.6. Strengthen the supporting role of standards in emerging industries. The standardizationBETTER COMMUNICATION ｜ GREATER VALUEproject in new industries will be carried out. According to the development trends and characteristics of industries, the plans for standards system in specific fields will be formulated to promote the development of key standards. National standards that lead the benign interaction and deep integration of the new-generation information technology and various industries will be developed. The standards system of industries regarding new materials will be continuously improved. Standards development in the field of biotechnology application will be strengthened. The development of standards for artificial intelligence medical devices, new biomedical materials, and new molecular diagnostic technologies will be accelerated to optimize the standards system for high-end and innovative medical devices. The development of digital technology standards will be strengthened to promote digital industrialization and industrial digitization. Standards on data security and data transaction will be developed to facilitate the development of the data element market.7. Promote the effective connection of standards between the upstream and downstream of the industrial chain. Standardization projects will be implemented to stabilize key industrial chains, and the standardization coordination of cross-industry and cross-domain industrial chains will be strengthened in key industries such as machinery, electronics and textiles. Standards development and application will be strengthened in key fields and products, as well as key links such as design, materials, technology, testing and application.8. Carry out a special standardization campaign on new infrastructure. More standards will be developed for the planning, design, construction, operation, and upgrading of new infrastructure such as industrial internet, internet of vehicles, energy internet and spatiotemporal information. Innovative infrastructure standards will be actively developed. Cities with a solid development foundation and prominent leading role will be selected to cultivate and apply new infrastructure standards. High-tech standardization pilot programs will be implemented to promote the development and application of integrated infrastructure standards.9. Implement the standardization improvement project on carbon peak and neutrality. The implementation plan for establishing and improving carbon peak and neutrality standards measurement system will be issued. The overall coordination role of standardization will be strengthened in various fields, and the national carbon peak and neutrality standardization group will be set up. The basic general standards for carbon peak will be improved, and mandatory national standards for energy consumption quotas in key industries and the energy efficiency of key energy-using products will be upgraded together with supporting standards for energy accounting, testing and certification, evaluation, and auditing. Verification standards for carbon emission accounting reports will be developed for regions and key industries, enterprises and products. Greenhouse gas emission standards for key industries and products will be developed. Standards on new power system will be developed, focusing on the clean and efficient use of wind power, photovoltaics, power transmission and distribution, energy storage, hydrogen energy, nuclear power and fossilenergy. Research will be conducted to develop standards for ecological carbon sinks, carbon capture, utilization and storage. The standardization pilot project of carbon peak and neutrality will be carried out. The construction and evaluation standards for green public institutions will be developed in different categories.10. Improve the standards system for ecosystem protection and restoration. Standards for ecological environment quality and risk management and control will be improved, and standards for ecological system environmental observation, environmental protection, and pollutant discharge will be developed and revised. The standards development in the fields of overall observation, protection and systematic restoration of mountains, rivers, forests, farmlands, lakes, grasslands, deserts and glaciers will be strengthened, and the standards system for meteorological guarantee services for ecological civilization, and the ecological protection and restoration of land and space will be established. Standards will be developed in areas such as biodiversity conservation and management, biosecurity assessment, ecological status monitoring and assessment, ecosystem stability assessment, ecological risk assessment and early warning, and ecosystem services. Green product standards will be improved, and the standards system for national parks, nature reserves, and ecological protection red lines will be constructed.11. Improve the standards level of economical and intensive utilization of natural resources. Unified technical standards for territorial and spatial planning and survey, registration, evaluation and monitoring of natural resource will be developed. More efforts are put into the development and revision of standards for the economical and intensive use of land resources and the management and classification of cultivated land resources. The standards development in the field of economical, intensive use and protection of water resources will be strengthened. Standards on mineral resource economical and intensive technologies, green exploration and green mining will be developed. Standards in the fields of seawater desalination and ocean energy utilization will be developed and revised.12. Carry out standardization actions for rural revitalization. The research into safety, quality, service and support standards for the entire agricultural industry chain will be conducted to build the standards system for modern agriculture. Standardization will be used to improve the rural living environment by strengthening the development of standards on rural garbage and sewage treatment, rural sanitary toilet construction and renovation, agricultural and rural infrastructure and public services, and improving related standards on construction, operation management, and supervision service. The standardization of rural governance will be promoted by developing and revising standards for rural security prevention and control. National agricultural standardization demonstration zones will be constructed, and the regional service and promotion platform for agricultural standardization will be established to consolidate and expand the standardization results of poverty alleviation.BETTER COMMUNICATION ｜ GREATER VALUE13. Improve the standards system for administrative management and social governance. Standardization actions on administrative management and social governance will be carried out to develop and revise standards in areas such as administrative licensing norms, government service improvement, national integrated government service platform, etc. The construction of standardization technical organizations on administrative management, government services and organizational governance will be promoted. Standardization pilot projects will be carried out in areas such as administrative management and government services, agency affairs services and community governance. National standards for credit system construction will be developed and revised to promote the interconnection and interoperability of credit information, and data sharing. Urban standardization actions will be taken to improve the standards system in key areas such as smart cities and sustainable urban development, and the construction of innovative cities with international standards will be strengthened. Urban standardization pilot projects will be implemented focusing on regional coordinated development, emergency response and grassroots community governance.14. Launch public safety standardization projects. An emergency management standards system will be built based on mandatory standards that adapt to all disasters and major emergencies. The development and revision of standards will be strengthened in areas such as social security, criminal law enforcement, anti-terrorism, disaster prevention and mitigation, comprehensive emergency management, pandemic prevention and treatment, food and material reserves, etc. Pilot projects for standards application will be launched. A unified and efficient public safety standardization coordination mechanism will be established to coordinate the work of multiple departments, regions and systems.15. Establish the standards system for basic public services. The development of standards for basic public services will be accelerated. The standardization of basic public services will be incorporated into regional development strategy. In the Yangtze River Delta region, the standardization management of basic public services will be fully implemented, and the standards system for basic public services will be established and improved. Standardization work such as standards development, application and evaluation will be carried out by grassroots service agencies, and a monitoring mechanism for the application of basic public service standards will be set up.16. Improve high-quality living standards. The standardization campaign for elderly care and housekeeping services will be carried out together with standardization pilot projects, to incorporate standardization into the leading action of improving the quality and expansion of domestic service industry. The publicity, implementation, supervision and inspection of mandatory national standards for service safety of nursing homes will be strengthened, and a unified national evaluation system for the service quality of elderly care institutions will be established. The development of health standards for infants and the elderly will be strengthened. The standards system will be completed in the areas of traditional Chinese medicine, sports, culture, tourism, radio, television and internet-based programs, cultural relics protection and utilization. Standards for vacation and leisure, rural tourism and homestays will be developed and applied, to support the digitization of cultural relics, archaeology, natural disaster prevention and other fields.17. Increase the participation in international standardization activities. The ability to participate in the development, conversion and application of international standards will be improved, and international standardization talents building will be strengthened to create a closer, mutually beneficial and win-win partnership on standards. China will fulfill the responsibilities and obligations as a member state of international standards organizations, actively participate in the formulation of strategic plans and organizational governance of international standards organizations such as ISO, IEC, ITU and CAC. China will also promote the development of international standards on carbon peak and neutrality, economical and intensive utilization of resources, smart cities, food safety, etc., and actively participate in international standardization activities for people’s well-being, gender equality, high-quality education, and cultural heritage protection.18. Deepen international standardization cooperation. Standards cooperation in the construction of free trade zones will be strengthened. Within the framework of the Regional Comprehensive Economic Partnership Agreement, the coordination of standards among members will be promoted. China will increase dialogues on standardization under the BRICS, APEC and other mechanisms, deepen exchanges with regional standard organizations, and strengthen standardization cooperation with Northeast Asia.19. Facilitate the compatibility of Chinese standards with international standards systems. A long-term working mechanism for the tracking and adoption of international standards will be established. The comparative analysis and adaptability verification of international standards and national standards in various fields will continue to be carried out to accelerate the adoption of advanced and applicable international standards. More foreign language versions of mandatory national standards will be published, and industries and local governments are encouraged to develop foreign language versions of standards as needed.20. Strengthen the standards connectivity in the Belt and Road Initiative. Standards cooperation will be incorporated in the Belt and Road cooperation documents. Universally applicable international standards will be applied in major projects and international scientific and technological cooperation. The national standards information platform for the BRI will be further completed to promote standard information sharing and cooperation, and mutual recognition of Chinese and foreign standards.21. Promote the coordinated development of domestic and international standardization. An international standardization working mechanism will be established, which is guided by the government, integrating production, education and research with enterprises as the main body. National standards will be gradually proposed and developed simultaneously with corresponding international standards, and the conversion and application of national standards and internationalBETTER COMMUNICATION ｜ GREATER VALUEstandards will be strengthened. Foreign-invested companies are encouraged to participate in standards development in accordance with the law. Enterprises, social organizations and research institutions are encouraged to actively participate in international professional standards organizations. And international professional standards organizations are supported to settle in China.22. Promote the coordinated development of government-led standards. The analysis of mandatory sectoral standards for safety production, public security and taxation will be made, and those comply with the scope of national standards will be transformed into national ones. The reform of voluntary national, sectoral and local standards will be implemented. The mechanisms for the management of major standard project development and revision, the adoption of association standards in voluntary national standards, and standard digitization will be improved with innovation.23. Strengthen the guidance of association standards and specifications. The opinions on the high-quality development of association standards will be formulated and implemented. Outstanding association standards and organizations developing association standards will be cultivated to promote the application of association standards. A series of national standards for association standardization good practice will be improved, and self-evaluation and self-declaration are encouraged. News media should play a role in the positive guidance and supervision of association standards.24. Improve the standardization capability of enterprises. The enterprise standards self-declaration disclosure and supervision system will play an effective role with improved public service platform to guide more companies to carry out self-declaration and disclosure through the platform. Third-party bodies are encouraged to publish the list of enterprise standard forerunners in the fields of consumer products, equipment manufacturing, emerging industries and services. Enterprises are encouraged to build a standard cooperation mechanism with research institutions and upstream and downstream industries. The work mechanism of standards comparison and compliance will be completed to promote the application of results.25. Promote the innovative development of local standardization. The local standards system structure will be optimized to improve the quality of standards supply. The standardization cooperation between provinces and ministries will be deepened to promote the in-depth reform of local standardization. Centering on the major national and regional strategies, standards will be jointly developed and applied in a coordinated way.26. Improve the coordination of quality infrastructure standards. Driven by standards, the integrated development of metrology, inspection and testing, certification and accreditation will be promoted, and the system for the integrated development of national quality infrastructure will be completed. The construction of national quality standard laboratories and national standard verification points will be promoted, and the construction of national technical standards innovation bases will be improved. The one-stop service platform for national quality infrastructure will be completed to provide full-chain technical solutions.27. Strengthen the implementation and supervision of standards. Advanced standards will be applied in the regulations and policies and the activities such as certification and accreditation, inspection and testing, government procurement and bidding. The statistical analysis of the implementation of mandatory national standards will be carried out. Policies and measures and information feedback mechanism for standards implementation will be optimized, and the third-party evaluation of standards implementation effect will be promoted. The supervision and random inspection of association standards and enterprise standards will be strengthened.28. Strengthen the construction of standardization technical support system. The construction of standardization technology organizations will be promoted in the development of new industries, new formats and new models. A liaison mechanism will be established to strengthen supervision, inspection and assessment, and promote the integration and optimization of standardization technical committees. A national digital standards library will be built to promote the construction of standards libraries at all levels.29. Support the development of standardization service industry. The standardization service market will be expanded to provide services such as retrieval and analysis, verification and evaluation, and pilot and demonstration. Government procurement of more standardization services will be encouraged to realize diversified service providers and methods. Surveys on the status quo of the industry will be conducted to study and propose the statistical plans and evaluation mechanisms for the industry.30. Strengthen the education and training of standardization talents. Standardization courses will be introduced in relevant majors, and the academic certificate + vocational skill grade certificate (1+X certificate) system will be promoted. Standardization talent training bases will be set up to increase the construction and personnel training of standardization majors at all levels.31. Improve incentive policies. Local governments should provide financial support for the implementation of important tasks in the Outline according to their financial situation. Commercial banks and financing guarantee companies are encouraged to develop standards financing and credit enhancement service products.32. Strengthen supervision and inspection. Local governments should incorporate standardization work into government performance evaluation. Standardization development evaluation pilot projects will be implemented in some regions, and the standardization statistical investigation system will be completed.33. Strengthen publicity and guidance. The significance, objectives and measures of implementing the Outline will be publicized in various forms. The interpretation of standardization policies will be released in time to improve the public’s awareness of standardization. Standardization publicity and popularization bases will be set up, and new media tools will be used to improve the accuracy and effectiveness of publicity.(Translated by CSP based on the Chinese version of the Action Plan)。

第34卷 第10期2011年10月计 算 机 学 报CH INESE JOURNA L OF COM PU TERSV ol.34N o.10Oct.2011收稿日期:2011-06-25;最终修改稿收到日期:2011-08-08.本课题得到国家自然科学基金项目(61063008)、云南省应用基础研究基金项目(2010CD025)、云南大学科学研究基金项目(2009F29Q)资助.欧阳志平,男,1985年生,硕士研究生,主要研究方向为数据挖掘.E -mail:chxoyzp@.王丽珍(通信作者),女,1962年生,博士,教授,主要研究领域为数据库、数据挖掘和计算机算法.E -m ail:lizhw ang2005@.陈红梅,女,1976年生,博士研究生,讲师,主要研究方向为数据挖掘.模糊对象的空间Co -location 模式挖掘研究欧阳志平 王丽珍 陈红梅(云南大学信息学院计算机科学与工程系 昆明 650091)摘 要 空间co -locatio n 模式表示的是空间对象的实例在一个相同的区域内频繁地空间并置.过去人们已经对确定及不确定数据的co -locatio n 模式挖掘问题进行了一些研究,但是针对模糊对象上进行的研究还没有.模糊对象在许多领域里都有着非常重要的应用,比如生物医学图像数据库和GIS.该文研究模糊对象的空间co -locatio n 模式挖掘问题.首先,定义模糊对象上空间co -locatio n 模式挖掘的相关概念,包括模糊参与率、模糊参与度等.其次,提出FB 算法挖掘模糊对象的co -location 模式.接着,提出了3种改进算法,包括剪枝对象、减少实例间连接、改进剪枝步,以提高挖掘性能、加快co -locatio n 规则的产生.最后通过大量的实验说明FB 算法及其改进算法的效果和效率.关键词 模糊对象;co -location 模式;空间数据挖掘;模糊参与率;减少连接中图法分类号T P 311 DOI 号:10.3724/SP.J.1016.2011.01947Mining Spatial Co -Location Patterns for Fuzzy ObjectsOUYANG Zh-i Ping WANG L-i Zhen CH EN H ong -M ei(Dep ar tment of Comp uter S cience and Eng ineering ,Sch ool of Inf or mation S cience and Engineering ,Yunnan Univ ersity ,K unming 650091)Abstract A spatial co -location pattern is a g roup of spatial objects w hose instances ar e frequently located in the sam e r eg io n.The mining co -location pattern pro blem for certain and uncertain data had been investigated in the past,but not for fuzzy o bjects.Fuzzy objects could be applied to many ar eas such as biom edical image databases,GIS and more.T his paper investigates the spatial co -location pattern m ining problem for fuzzy objects.Firstly ,it defines the r elated concepts of spatial co -location patterns mining on fuzzy objects,including fuzzy par ticipatio n ratio,fuzzy par -ticipation index ,etc.Secondly ,this paper proposes an FB algorithm to mine co -location patterns fro m fuzzy objects.Then,three kinds of the im pr oved alg orithm s,the pr uning objects,reducing of the operation joining betw een spatial instances and optimizing the pruning steps,are put for -w ard so as to impro ve the mining performance and accelerate the co -location rule g eneration.Finally,by ex tensive ex periments,the efficiency and effectiveness of the algorithms are verified.Keywords fuzzy o bjects;co -location patterns;spatial data mining;fuzzy participation index;reducing jo ining1 引 言空间co -lo catio n 模式代表了一组空间对象的子集,它们的实例在空间中频繁地关联.挖掘空间co -location 模式就是在空间数据库中发现和挖掘空间对象之间的关联关系.例如,西尼罗河病毒通常发生在蚊子泛滥、饲养家禽的区域;植物学家们发现/半湿润常绿阔叶林0生长的地方80%会有/兰类0植物生长[1].在现实世界中,模糊对象无处不在,比如/漂亮的女人0、/高大的树0等,同时模糊对象在许多应用中也起着十分重要的作用,目前对模糊对象的研究范围十分广泛,但对co-location模式的研究还没有,针对这一情况,本文研究模糊对象上的co-lo catio n 模式挖掘问题.本文第2节为相关工作;第3节是相关定义及性质;第4、5节为算法和实验;第6节为结论.2相关工作空间co-lo catio n模式挖掘问题是空间数据挖掘领域的一个重要研究方向,人们对确定数据上的co-location模式挖掘问题进行了深入的研究,并提出了很多算法,比如join_based算法[2]、partia-l join 算法[3]、join-less算法[4]、CPI-tree算法[5]、order-clique-based算法[6]等.在文献[2]中,给出了co-location模式挖掘相关的一些定义,包括邻近关系、空间co-location模式、行实例、表实例、参与率、参与度以及co-lo catio n规则和条件概率等.近年来,对不确定数据上的co-location模式挖掘的研究也越来越多,文献[7]提出了不确定集上的U Join_based算法.文献[8]研究了从区间数据表示的不确定对象中挖掘co-location模式.虽然目前空间co-location模式挖掘算法很多,但对模糊对象的co-location挖掘算法还未见报道.模糊对象[9]的研究目前主要集中在模糊对象的建模上,比如基本的类型和操作模型等.文献[10-13]在GIS中对模糊对象做了大量的研究.文献[14]研究模糊对象的K NN查找问题,提出了A K NN和R K NN算法.3相关定义及性质本部分首先对模糊对象、模糊概率阈值、空间距离、模糊参与率以及模糊参与度进行定义,其次给出本文定义的模糊参与率及参与度所满足的一个性质.3.1相关定义定义1(模糊对象).本文中的模糊对象,表示为一个空间中离散点的集合,定义如下:A= {〈a,L(a)〉|L(a)>0},其中A表示模糊对象,a表示实例,L(a)表示实例a属于模糊对象A的隶属度.图1所示为一个模糊对象A,实例a1、a2属于A的隶属度分别为016,0101.图1模糊对象从图1中可以看出,实例a2属于模糊对象A 的隶属度非常低,对于这样的实例,我们希望计算co-location模式时把它排除,因为在现实生活中,我们认为隶属度非常低的实例对于模糊对象来说没有多大意义.例如老年人,假设80岁隶属度为018,50岁为0102,对于50岁,我们认为它对于老年人这个对象贡献相当少,所以在有些应用中可以把它排除,而只关注那些符合实际应用的期望隶属度值条件的实例,于是我们有以下定义.定义2.给定一个用户自定义的概率阈值f_threshold,称为模糊度阈值,集合A f_thr eshold={a| L(a)E f_thr eshold},表示满足用户自定义模糊度阈值的实例集.如图1中,A012={a1}.定义3(空间邻近关系R).设a,b分别为模糊对象A,B的实例,它们之间的距离用欧几里德距离计算,表示为d(a,b)=+a-b+.由此可以定义空间邻近关系R:若两个实例之间的欧几里德距离小于等于阈值dis_thr eshold,即d(a,b)F dis_thr esh-old,则表示它们邻近.当两个空间实例a和b之间满足空间邻近关系R时,称这两个空间实例R邻近,记为R(a,b),并在图中用线段连接它们(如图2所示).图2模糊对象参与度一个空间co-location模式表示的是一组空间对象的集合.co-location模式的长度称为此co-location 模式的阶,即co-location模式里空间对象的个数.例如co-location模式c={A,B,C},则称模式c是3阶co-location模式.设有空间实例集I={i1,i2,,,i l},如果有{R(i j,i k)|1F j F l,1F k F l},则称I是一个团1948计算机学报2011年(clique ).如果团I 包含了co -locatio n 模式c 中的所有对象,并且I 没有任何一个子集可以包含c 中的所有对象,那么I 是co -location 模式c 的一个行实例(称为co -lo cation 实例).co -lo cation 模式c 的所有行实例的集合称为表实例,记为table _instance(c).例如图2中,co -location {A,B,C}的表实例为{{a 2,b 1,c 1},{a 2,b 3,c 1}}.定义4(模糊参与率). 设f i 为某个空间模糊对象,f i 在k 阶的co _location 模式c 中的模糊参与率表示为FP R (c,f i ),它是f i 的实例在空间co -location 模式c 的表实例中不重复出现的实例的隶属度之和与f i 中总实例个数的比率,公式如下:FP R (c,f i )=E 1@L (a)|table _instance({f i })|,其中a I Ffi(table _instance(c)),而F 是关系投影操作.定义5(模糊参与度).模糊对象的空间co -location 模式c ={f 1,f k }的模糊参与度表示为FPI (c),是模式中所有空间对象FP R 值中的最小值,公式表示如下:F PI (c)=m in ki =1{F PR(c,f i )}.设min _p r ev 是用户给定的最小参与度阈值,当FPI (c)E min _p r ev 时,称模糊对象的co -lo catio n 模式c 是频繁的.例1. 图2所示为3个空间模糊对象A ={(a 1,012),(a 2,014),(a 3,016),(a 4,018)},B ={(b 1,013),(b 2,016),(b 3,018)},C ={(c 1,011),(c 2,017)},R 关系用连线表示,假设min _p r ev =012.则在co -location 模式c ={A ,C}中,FP R (c,A )=(1@014+1@018)/4=013,而FP R (c,C)=(1@011)/2=0105,则F PI (c )=min {013,0105}=0105,co -lo cation 模式c 为非频繁模式.定义4说明:假设模糊对象A 有4个实例,图3表示的是A 的一个实例a 1的模糊参与率与其隶属度的关系.横坐标表示a 1属于对象A 的隶属度值,纵坐标表示模糊参与率值,从图中可以看出,当实例a 1的隶属度增加时,它的模糊参与率也会相应增加,因为a 1的模糊参与率值为a 1的隶属度与对象A 的总实例个数的比率.当隶属度为0时表示a 1不属于对象A ,这时模糊参与率自然为0,隶属度为1时表示a 1完全属于对象A,这时它的模糊参与率与传统定义的值一样,为0125.图3 定义4说明3.2 性 质模糊参与率与参与度满足如下的一个引理和定理.引理1.模糊参与率(FPR)和模糊参与度(FPI )随着co -location 模式阶的增大单调递减.证明. 假设某个模糊空间对象的实例包含在co -location 模式c 的实例中,那么当有co -location 模式c 1A c,这个模糊空间对象的实例也一定包含在模式c 1的实例中,反之则不然,所以模糊参与率是单调递减的.由于模式的模糊参与度取它包含的对象中最小参与率值,当模式的阶增大时,由于模糊参与率是递减的,所以co -location 模式的模糊参与度也是单调递减的.定理1. 如果k 阶co -locatio n 模式c 是频繁的,那么它的任意k -1阶的子co -lo catio n 模式也是频繁的.证明. 根据引理1,k 阶co -location 模式c 的模糊参与度,要小于其k -1阶子co -location 模式,所以若k 阶co -location 模式c 是频繁的,则k -1阶的子co -location 模式肯定也是频繁的.利用定理1可以对候选模式进行剪枝处理,具体的剪枝过程见414节.4 模糊对象co -location 模式挖掘算法首先给出一个模糊对象的co -locatio n 挖掘的基本算法)))FB 算法,接着在其基础上提出3种改进算法,包括剪枝模糊对象、减少实例间的连接、优化剪枝步.算法中模糊对象的实例按照实例的模糊度非递增排序.4.1 FB 算法FB 算法采用的是经典jo in_based 算法的思想,循环执行以下4个步骤:(1)产生候选co -location 模式(包括连接步和剪枝步);(2)产生候选co -location 模式的表实例;(3)剪枝(利用用户自定义的参与度阈值min _p r ev 进行剪枝);(4)产生co -location 规则.具体过程如算法1所示.194910期欧阳志平等:模糊对象的空间Co -location 模式挖掘研究算法1.FB算法.输入:空间模糊对象集SF,空间实例集S FI,参与度阈值min_p rev,条件概率阈值min_conf,距离阈值dis_thr esho ld,模糊度阈值f_thr eshold,满足模糊度阈值条件的模糊对象与实例的集合F,FI 输出:co-locatio n规则集FP变量:k:co-lo cat ion模式的阶,C k:k阶co-locatio n候选模式集,T k:候选模式C k的表实例集,P k:k阶co-locatio n频繁集,R k:k阶co-location规则集步骤:1.F,FI=g en_fdata(SF,SF I,f_threshold);2.P1=F,FP=ù;3.fo r(k=2;P k-1Xù;k++)do311.C k=gen_candidate_co-lo cat ion(k,P k-1);312.T k=g en_table_instance(C k,T k-1);313.P k=sel_prev_co-lo cation(C k,T k,min_p r ev);314.R k=sel_co-lo cation_rule(P k,T k,min_conf);315.FP←FP G R k;4.return FP.步1是根据文中定义的模糊度阈值得到满足条件的模糊对象和实例集;步2为初始化;步3迭代地生成频繁co-location模式集和规则集,其中,步311为生成k阶co-location候选模式集,步312为生成k阶co-location候选模式的表实例集,步313生成k 阶频繁co-location模式集,步314生成k阶co-location 规则集;步4返回结果.4.2剪枝模糊对象由于co-lo catio n挖掘算法需要对模糊对象的大量实例之间进行距离计算以及连接操作,因此我们应当尽可能剪掉那些不可能存在于co-lo catio n 模式中的模糊对象.基于以上思考,论文提出一种有效的剪枝规则.定理2.对于一个模糊对象A,若它最大模糊参与率值小于给定的最小参与度阈值,则对象A不可能存在于任意的频繁co-lo catio n模式中.证明.反证法.假设模糊对象A存在于某个频繁的co-location模式c中,则我们可以得到FPR(c,A)E min_p rev.模糊对象A在c中的最大参与率为它的所有实例均在co-lo catio n模式c的表实例中,根据模糊参与率的定义,最大模糊参与率值等于模糊对象A的所有实例的隶属度之和与对象A的实例数目的比率,由定理条件可知,它小于给定的最小支持度阈值,这时可以得到F PR(c,A)< min_p rev,与假设矛盾,所以对象A不可能存在于任意的co-location模式中.例2.图2中,设m in_p r ev=016.假设模糊对象C的所有实例均在co-lo cation模式的行实例中, C的最大参与率值等于011/2+017/2=014<016,由定理2可知,C不可能存在于任意的co-location 模式中,这时把对象C剪枝掉.利用定理2,可以在模式挖掘前对对象进行初步剪枝,降低算法的时间复杂性.具体见算法2,其中步骤1为计算每一个模糊对象的最大参与率.算法2.剪枝模糊对象算法.输入:模糊对象集F,实例集FI,参与度阈值min_ p rev,对象的最大参与率T_P R输出:剪枝模糊对象后的F,FI步骤:1.fo r all fuzzy o bject f I F do111.计算f的最大参与率T_PR112.if T_P R<min_p r ev thenF=F-{f};FI=F I-{a}(其中a I f);2.retur n F,FI.4.3减少实例间连接尽管通过312节中的改进算法可以减少实例之间的计算量,但co-lo catio n模式挖掘过程中仍然有大量实例之间的连接操作,相当耗时.在FB算法基于参与度的剪枝过程中,首先要生成候选模式的表实例,再基于参与度阈值来对候选模式进行剪枝.我们在研究中发现,可以在表实例生成最开始阶段就对某些不可能满足参与度阈值的候选模式进行剪枝,避免表实例之间大量不必要的连接操作,大大提高算法的效率.下面给出定理3,它是减少实例间连接算法的依据.定理3.在co-location模式c中,假设模糊对象A I c,如果A在c的表实例中的实例满足m ax{L(a)}<min_p rev,其中a是对象A的实例,则co-location模式c可以被剪枝掉.证明.因为FP R(c,A)FE ni=1L(a i)|table_instance(A)|F n@max{L(a i)}|table_instance(A)|=n@max{L(a i)}n=max{L(a i)},所以当m ax{L(a)}<m in_p r ev时co-location模式可以被剪枝(假设模糊对象A的实例数是n).例3.图2中,考虑2阶co-location模式c= {A,B},每个对象的实例序按隶属度非递增进行排序,假设min_p rev=017.对象A与对象B实例的R1950计算机学报2011年关系中,a 4与B 的实例不存在R 关系,与B 存在R 关系的最大隶属度实例为a 3,而a 3的隶属度为016<min _p r ev ,所以模式c ={A ,B}可以被提前剪枝,这样就不用计算实例a 2,a 1与对象B 的实例之间的关系.从定理3可知,只要模式中任何一个对象,它在模式表实例中的实例的最大模糊度小于参与度阈值,则可以对该模式进行剪枝.于是可以得到减少实例间连接的改进算法,具体见算法3,其中变量max _f lag 为判断是否最大模糊度实例条件,因为对象的实例顺序已经按照实例模糊度非递增排序,所以在实例连接过程中,第一次实例间的连接即为模式每个对象具有最大模糊度的实例间的连接.算法3. 减少实例间的连接算法.输入:频繁k 阶模式p 1k ,p 2k ,模式p 1k ,p 2k 的表实例t 1k ,t 2k ,p 1k ,p 2k 连接生成的k +1阶候选模式c 3k +1输出:候选模式c 3k +1的表实例集t 3k +1步骤:1.max _f lag =true ;2.fo r (i =1;i F t 1k .count;i ++)2.1.for (j =1;j F t 2k .count;j ++)2.1.1.if ((t 1k [i].instance 1=t 2k [j ].instance 1,t 1k [i].instance k -1=t 2k [j ].instance k -1)and (t 1k [i].i nstance k ,t 2k [j ].ins tance k )I R )then2.1.1.1.if (max _f lag =tr ue and(t 1k [i].instance <min _p r ev or t 2k [j ].ins tance <min _p r ev))thengo to loop1;2.1.1.2.elset 3k +1=t 3k +1G {t 1k [i].instance joint 2k [j ].instance};2.1.1.3.max _f alg =false ;3.loo p1:r eturn t 3k +1.4.4 优化剪枝步这一小节,本文将提出一种新的改进算法,对FB 算法中生成候选模式过程中的剪枝步进行优化,在FB 算法中剪枝步的做法如下:删除所有C k 中这样的候选模式c,c 的k -1阶子集不是频繁的.本文根据第2部分中的定理1提出一种新的剪枝策略,它不用去比较c 的每一个k -1阶子集是否频繁,就可以对候选模式进行剪枝.实验和理论分析表明,新的剪枝步策略要比传统的策略具有更好的时间复杂性.对候选k 阶模式C k 增加一个相应的计数数组.该计数数组产生的过程和功能简单描述如下:当频繁模式进行连接产生候选co -location 模式时,增加一个计数数组.连接过程中每产生一个候选co -location 模式,数组就相应地增加一位,用来存储该候选模式的频繁k -1阶子模式的个数.由定理1可知,每一个k 阶候选模式,其(k -1)阶子模式都是频繁的且为k 个,所以最后可以根据数组相应位的值来判断对应的候选模式是否频繁.以下是具体的处理方法和过程.图4 优化剪枝步算法图4(a )中,假设已得到2阶频繁模式为P 2={A B,AC,AD,B C,BD},这时基于P 2连接产生3阶候选模式过程为:首先增加一个计数数组CR ,对于P 2中的模式,计算A B,由于C 3为空,所以对数组CR 不做任何处理.接着A B 开始与P 2中其它频繁模式连接产生候选模式,A B 和AC 连接产生候选195110期欧阳志平等:模糊对象的空间Co -location 模式挖掘研究模式A BC ,这时相应地将数组CR 增加1位,并将其值加1.A B 与A D 连接产生候选模式AB D ,数组CR 增加1位,并将其值加1.A B 连接完成,开始考虑A C,先扫描3阶候选模式,由于A C 是A BC 的子集,所以CR[1]加1.接着A C 与A D 进行连接产生候选模式A CD ,CR [3]加1,依此循环,直到P 2中所有频繁模式连接完成,详细过程见图4(b).最后C 3={ABC,A BD,ACD,BCD},数组CR ={3,3,2,2},由于A CD 和BCD 计数位不为3,所以对其进行剪枝.对于每一个k 阶候选模式,若其频繁子模式个数不为k,则对其进行剪枝.通过证明,新的剪枝步策略比传统的具有更好的时间复杂性,以下给出证明.证明. 假设C k 中有m 个模式,P k -1中有n 个频繁(k -1)阶模式,在传统剪枝步中,每一个k 阶模式有k 个(k -1)阶子模式需要与P k -1进行比较,其时间复杂度为k @m @n,而在新的剪枝歩策略中,只需对P k -1中每一个模式扫描一遍C k ,其时间复杂度小于m @n,由此可知新策略的时间复杂度更优.改进算法的伪代码见算法4.算法4. 优化剪枝步算法.输入:k 阶频繁模式集P k ,存储对应的候选模式的频繁子集数目CR[]输出:k +1阶候选模式集C k +1步骤:1.fo r (i =1;i F P k .count;i ++)1.1.for all C k +1[x ]I C k +1/*判断频繁模式是否为k +1阶候选模式的子集*/if (P k [i]中对象都在C k +1[x ]中)then CR[x ]=CR[x ]+1;1.2.for (j =i +1;j F C k .count;j ++)/*k 阶频繁模式间连接*/if (P k [i].obj ect 1=P k [j ].obj ect 1,P k [i].obj ect k -1=P k [j ].obj ect k -1)thenC k +1=C k +1G {P k [i]jo in P k [j ]};2.fo r all C k +1[y]I C k +1/*判断k +1阶候选模式是否频繁*/if CR[y ]X k thenremov e C k +1[y ]fr om C k +1;3.return C k +1.5 实验与分析在本节中,本文做了大量实验来验证所提出的FB 算法和改进算法的有效性,并用文中提出的算法与传统算法的挖掘结果进行了实验比较.所有算法均采用C#编写,并在AMD A thlon 118GH z CPU 和512M B memor y 的计算机上运行.实验所采用的实例数据均是随机产生的并均匀分布在100@100空间里,模糊对象的数目为10,模糊度的值从0到1,也是随机产生的.表1给出了实验的参数以及默认值.表1 实验数据的参数说明参数默认值实例数目500参与度阈值015距离阈值15模糊度阈值5.1 FB 算法与其改进算法的性能比较在这一小节中,本文将对模糊对象co -location 挖掘算法(FB)与其改进算法进行比较,改进算法包括剪枝对象算法(PO)、剪枝对象基础上的减少实例间连接的算法(PO_RI)、剪枝对象且减少实例间连接基础上的优化剪枝步算法(PO_RI_PC).5.1.1 实例数目对算法的影响首先考查实例数目对算法的影响,实例数目从500增加到5000.从图5可以看出,随着实例数目增加,所有算法的运行时间都增加,因为随着实例数目的增加,算法实例间距离计算、连接操作也会增加.PO_RI 和PO_RI _PC 算法由于采用了减少实例间的连接改进算法,故随着实例数目的增加,它的处理时间上升的幅度比较小,这也说明了减少实例间连接改进算法的高效性.图5 实例数目对算法的影响5.1.2 参与度阈值对算法的影响下面研究参与度阈值对算法的影响,参与度阈值的变化从018到012.从图6中可以看出,FB 算法和PO 算法的运行时间随着参与度阈值的降低而急剧上升,而PO_RI 算法和PO_RI_PC 算法的运行时间一直保持平稳,这是因为随着参与度阈值的降低,更多的co -lo cation 模式满足参与度阈值条件,1952计 算 机 学 报2011年使得FB 和PO 算法运行时间急剧上升,而PO_RI 和PO_RI_PC 算法由于采用了减少实例间的连接改进算法,所以运行时间一直较平稳.从图中还可以看出当参与度阈值在014到012区间时,FB 和PO 算法运行时间基本一样,这是因为PO 算法采用的是剪枝模糊对象改进算法,它依赖于参与度阈值,当自定义的参与度阈值非常低时,算法很难剪枝掉很多的模糊对象,所以两个算法的运行时间相差不大.图6 参与度阈值对算法的影响5.1.3 距离阈值对算法的影响接下来考虑距离阈值对算法运行时间的影响,距离阈值变化从10~40.从图7中可以看出,类似于实例数目对算法的影响,随着距离阈值增大,FB 和PO 算法运行时间快速上升,而PO_RI 和PO_RI_PC 算法上升较平稳.图7 距离阈值对算法的影响5.1.4 模糊度阈值对算法的影响本组实验的最后考虑模糊度阈值对算法的影响,模糊度阈值从0到015变化.从图8中可以看出,当模糊度从0到012时,4个算法的运行时间均上升,而在012以后,4个算法的运行时间又开始下降,这是因为把模糊对象的那些不满足模糊度阈值的,具有低模糊度的实例剪去后,剪枝掉实例的对象的参与率值会增大,这时满足参与度阈值条件的模式数目会增多,使得算法运行时间上升.但是随着模糊度阈值的不断增大,要剪去对象的实例也越多,这就意味着一个对象具有的实例数目越来越少,就会减少模式产生的数目,使得算法的运行时间下降.图8 模糊度阈值对算法的影响5.2 模糊挖掘算法与传统算法的比较在这一小节,本文用模糊对象co -lo cation 挖掘算法,与传统挖掘算法的结果进行实验比较.这里传统算法采用的是co -lo cation 挖掘算法中最为经典的join_based 算法[2].考虑实例个数、参与度阈值、距离阈值、模糊度阈值对两种算法的影响,其中模糊对象co -location 模式挖掘算法采用的是PO_RI_PC 算法.5.2.1 实例个数对算法的影响首先,考虑实例个数对两种不同算法的影响,实例数目从500增加到5000.从图9中可以看出,两种算法的模式数目都随着实例数目的增加而增加.Join_based 算法产生的模式数目远大于PO_RI_PC 算法,这是因为PO_RI_PC 算法模式参与度的计算是根据实例的模糊度来计算的,而传统算法中没有区分实例之间模糊度的差别,这导致了后者的参与度要高于前者,在相同参与度阈值条件下,后者频繁模式数目要明显高于前者.图9 实例个数对算法的影响5.2.2 参与度阈值对算法的影响图10 参与度阈值对算法的影响接下来考虑参与度阈值对算法的影响,阈值变化从018到012.从图10中可以看出两种算法的模195310期欧阳志平等:模糊对象的空间Co -location 模式挖掘研究式数目随着阈值降低都在增加,在低参与度阈值条件下,比如阈值为012,两者的频繁模式数目比较接近.5.2.3 距离阈值对算法的影响接下来考虑距离阈值对算法的影响,距离阈值的变化从10~40.从图11中可以看出,两种算法的模式数目都上升,但PO _RI _PC 算法上升较慢,因为限制频繁模式个数的因素还有参与度.图11 距离阈值对算法的影响5.2.4 模糊度阈值对算法的影响最后来看模糊度阈值对算法的影响,模糊度阈值的变化从0~015.从图12中可以看出,Jo in_based 算法的模式个数保持不变,因为模糊度没有参与模式参与度的计算中.而PO_RI_PC 算法生成的模式个数开始上升,后来出现下降,原因可以见实验51114节中的分析.图12 模糊度阈值对算法的影响6 结 论尽管空间co -lo cation 模式挖掘是一种非常有价值的空间挖掘,而且模糊对象也经常出现在许多重要的应用中,但是目前对于模糊对象的co -location 模式挖掘的研究还未见报道.本文针对模糊对象的空间co -location 模式挖掘问题,提出了一种基本挖掘算法)))FB 算法,为了提高算法的挖掘效率,文中提出了3种改进算法,包括剪枝对象、减少实例间连接、改进剪枝步.通过大量的实验表明,本文提出的算法及改进算法是非常有效的.下一步的工作将在此论文的基础上,考虑模糊度阈值在一个范围内变化时的空间co -location 模式挖掘问题.参考文献[1]W ang L-i Zhen ,Zhou L-i H ua,Ch en H ong -M ei et al.T he Principle and Applications of Data W arehous e and Data M ining.2nd Edition.Beijing:Science Press,2009(in Chin ese)(王丽珍,周丽华,陈红梅等.数据仓库与数据挖掘原理及应用.第2版.北京:科学出版社,2009)[2]H uang Y,Sh ekh ar S,Xiong H.Dis covering colocation patterns from spatial data s ets:A g eneral approach.IE EE T ran saction s on Kn owledge and Data E ngineering,2004,16(12):1472-1485[3]Yoo J S,S hekhar S.A partial join approach for min ing co -location patterns//Proceeding s of the ACM International Sym posium on Advances in Geographic Information S ystem s (ACM GIS).Washin gton,US A,2004:241-249[4]Yoo J S ,Shekhar S,Celik M.A join -less approach for co -location pattern min ing:A s ummary of results//Proceedin gs of the IEEE In ternational Conference on Data M in ing (ICDM ).H ouston,USA,2005:813-816[5]W ang L-i Zh en,Bao Yu -Zhen,Lu J,Yip J.A new join -les s approach for co -location pattern minin g//Proceedings of the IEEE 8th International Conference on Com puter an d Infor -mation Techn ology (CIT 2008).Sydney,Au stralia,2008:197-202[6]W ang L-i Zh en,Zh ou L-i Hua,L u J.Yip J.An ord er -clique -based approach for mining max imal co rmation Sciences,2009,179(19):3370-3382[7]Lu Ye,W an g L-i Zh en,Zhang Xiao -Feng.M inin g frequent Co -location patterns from u ncertain data.Journ al of Fron -tiers of Computer S cien ce and Techn ology,2009,3(6):656-664(in Chinese)(陆叶,王丽珍,张晓峰.从不确定数据集中挖掘频繁Co -location 模式.计算机科学与探索,2009,3(6):656-664)[8]W ang L-i Zh en,Ch en H ong -M ei,Zhao L-i H ong et al.Eff-i ciently mining co -location rules on interval data//Proceedin gs of th e 6th Internation al Conferen ce on Advan ced Data M in ingand Application s (ADM A 2010).Chongqing,China,2010:477-488[9]Zadeh L.Fuzz y s rm ation and Control,1965,8(3):338-353[10]Altman D.Fu zzy set th eoretic approaches for han dling im -precision in spatial an aly sis.International J ou rnal of Geo -graphical In formation Science,1994,8(3):271-289[11]Schneider M.Fuzzy topological predicates,their properties,and their integration into qu ery languages //Pr oceedings of the ACM International Symposiu m on Advances in Geo -graphic Inform ation Systems (ACM GIS ).New York,U SA,2001:9-141954计 算 机 学 报2011年。

基于多特征描述的乳腺癌肿瘤病理自动分级摘要：为了辅助病理医生快速高效诊断乳腺癌并提供乳腺癌预后信息，提出一种计算机辅助乳腺癌肿瘤病理自动分级方法。

该方法使用深度卷积神经网络和滑动窗口自动检测病理图像中的细胞；随后综合运用基于稀疏非负矩阵分解的颜色分离、前景标记的分水岭算法以及椭圆拟合得到每个细胞的轮廓。

基于检测到的细胞和拟合出的细胞轮廓，提取出肿瘤的组织结构特征和上皮细胞的纹理形状特征等共203维的特征，运用这些特征训练支持向量机分类器（SVM），实现对病理组织图像自动分级。

17位患者的49张H&E染色的乳腺癌病理组织图像自动分级的100次十折交叉检验评估结果表明：基于病理图像的细胞形状特征与组织的空间结构特征对病理图像的高、中、低分化等级分类整体准确率为90.20%；同时对高、中、低各分化等级的区分准确率分别为92.87%、82.88%、93.61%。

相比使用单一结构特征或者纹理特征的方法，所提方法具有更高的准确率，能准确地对病理组织图像中肿瘤的高级和低级分化程度自动分级，且各分级之间的准确率差异较小。

关键词：乳腺癌；组织病理图像；自动英文无病理病理分级；计算机辅助预后分析中图分类号：TP391.41；TP182文献标志码：AAbstract：In order to assist in the fast and efficient diagnosis of breast cancer and provide the prognosis information for pathologists，a computeraided diagnosis approach for automatically grading breast pathological images was proposed. In the proposed algorithm，cells of pathological images were first automatically detected by deep convolutional neural network and sliding window. Then，the algorithms of color separation based on sparse nonnegative matrix factorization，marker controlled watershed，and ellipse fitting were integrated to get the boundary of each cell. A total of203dimensional imagederived features，including architectural features of tumor，texture and shape features of epithelial cells were extracted from the pathological images based on the detected cells and the fitted boundary. A Support Vector Machine （SVM）classifier was trained by using the extracted features to realize the automated grading of pathological images. In order to verify the proposed algorithm，a total of 49 Hematoxylin & Eosin （H&E）stained breast pathological images obtained from 17 patients were considered. The experimentalresults show that，for 100 tenfold crossvalidation trials，the features with the cell shape and the spatial structure of organization of pathological image set successfully distinguish test samples of low，intermediate and high grades with classification accuracy of 90.20%. Moreover，the proposed algorithm is able to distinguish high grade，intermediate grade，and low grade patients with accuracy of 92.87%，82.88% and 93.61%，respectively. Compared with the methods only using texture feature or architectural feature，the proposed algorithm has a higher accuracy. The proposed algorithm can accurately distinguish the grade of tumor for pathological images and the difference of accuracy between grades is small.英文关键词Key words：breast cancer；pathological image；automated pathological grading；computeraided prognosis 0引言美国癌症协会2015年发布的数据显示，预计2015年美国女性乳腺癌新增病例仍居女性所有癌症病例的首位，占所有新增癌症病例的29%[1]。

测绘工程专业英语1. Introduction测绘工程是一门综合性较强的学科，需要使用专业的英语术语来进行交流和沟通。

本文将介绍一些与测绘工程相关的常用英语词汇和表达，以帮助读者更好地理解和运用专业英语。

2. Surveying and Mapping 测绘2.1 Surveying 规划测量•Geodetic surveying大地测量•Cadastral surveying (land) 地籍测绘•Engineering surveying工程测量•Topographic surveying地形测量•Control surveying控制测量•Hydrographic surveying水文测量•Deformation monitoring变形监测2.2 Mapping 制图•Cartography制图学•Digital mapping数字制图•Geographic information system (GIS)地理信息系统•Remote sensing遥感•Image interpretation图像解译•Orthophoto正射影像•Contour lines等高线•Land cover classification土地覆盖分类3. Instruments and Equipment 仪器设备•Theodolite经纬仪•Total station全站仪•GPS receiver GPS接收机•Levels水准仪•Laser scanner激光扫描仪•Data collector数据采集器•GIS software GIS软件•Mapping software制图软件4. Measurements 测量数据•Distance measurement距离测量•Angle measurement角度测量•Coordinate measurement坐标测量•Elevation measurement高程测量•Height measurement高度测量•Area measurement面积测量•Volume measurement体积测量5. Data Analysis 数据分析•Data processing数据处理•Error analysis误差分析•Statistical analysis统计分析•Spatial analysis空间分析•Data visualization数据可视化•Interpolation插值•Extrapolation外推•Regression analysis回归分析6. Project Management 项目管理•Survey project planning测量项目规划•Budget estimation预算估算•Resource allocation资源分配•Scheduling进度安排•Quality control质量控制•Risk assessment风险评估•Project documentation项目文档7. Conclusion本文介绍了一些与测绘工程专业相关的英语词汇和表达，涵盖了测绘、制图、仪器设备、测量数据、数据分析和项目管理等领域。

附录：并联位移机器人的设计Jacques M.HERVEECELE CENTRALE PARIS92295 CHATENAY MALABRY CEDEXFRANCE摘要摘要::本文目的是对偶具有人性化机器人的应用做一个完全的介绍,并将着重讨论并行机器人特别是那些能够进行空间平移的机器人。

在许多工业的应用过程中这种机器人被证明其末端执行器在空间上的定位是没必要的。

这个方法的优点是我们能系统地导出能预期得到位移子群的所有运动学链。

因此，我们调查了机器人的整个家族。

位移子群的所有运动学链。

因此，我们调查了机器人的整个家族。

T-STAR T-STAR 机器人现在就是一台工作装置。

而H-ROBOT,PRISM-ROBOT 是新的可能的机器人。

这些机器人能满足现代生产快节奏工作中价格低以及符合挑选的工作环境，如选料、快节奏工作中价格低以及符合挑选的工作环境，如选料、安排、包装、装配等发日益增长的安排、包装、装配等发日益增长的需求。

需求。

关键词：运动学：运动学,,并行机器人并行机器人引言群论可以运用于一系列位移当中。

根据这个理论，如果我们能够证明群群论可以运用于一系列位移当中。

根据这个理论，如果我们能够证明群{D}{D}{D}包含所有的包含所有的可能的位移，那么那么{D}{D}{D}就具有群结构。

就具有群结构。

就具有群结构。

刚体的最显著运动是由群刚体的最显著运动是由群刚体的最显著运动是由群{D}{D}{D}表现出来的。

表现出来的。

这方法导致机械装置的分类机械装置的分类 [1] [1]。

建立这样的一个分类的主要的步骤是将位移群的所有子群导出。

这能通过检验所有具有旋转和平移特性的通过检验所有具有旋转和平移特性的[2][2][2]产品直接推理出。

产品直接推理出。

然而然而,,一个更有效的方法存在于假设群论设群论[3],[4][3],[4][3],[4]中。

假设群论是在取决于许多有限实参数的全纯映射的基础上定义的。

位移中。

WalkField 软件在南极测绘中的应用王连仲1,孔繁宇2,朱　虹3(1.极地测绘科学国家测绘局重点实验室,黑龙江哈尔滨150025;2.哈尔滨市信息化建设项目管理中心,黑龙江哈尔滨150001;3.国家测绘局第四航测遥感院,海南海口570203)摘　要:介绍了具有自主知识产权的GIS 软件—WalkField ,及其在南极测绘中的应用情况,阐明了国产软件的优势,并介绍了应用此软件生成地形图的方法。

同时,文中对南极地图投影进行了阐述,对于南极地区大比例尺地形图,采用高斯—克吕格投影优于其它投影方式,具有良好的效果。

关键词:WalkField ;GIS 软件;南极;地图投影中图分类号:P28 文献标识码:A 文章编号:1006-7949(2007)03-0001-05Application of W alkField soft w are in antarctic surveying and m appingWAN G Lian 2zhong 1,KON G Fan 2yu 2,ZHU hong3(1.K ey Laboratory of Pole Surveying and Mapping Science ,SBSM ,Harbin 150025,China ;2.Harbin City Project Management Center of Information Technology Constructions ,Harbin 150001,China ;3.4th Institute of Photogrammetry and Remote State Bu 2reau of Surveying and Mapping ,Haikou 570203,China )Abstract :It introduces the copyright —WalkField software of GIS of the independent intelligent ,which has been applied in South Pole surveying and mapping circumstance.It presents the advantages of the domestic soft 2ware ,and the methods of this software for production map.At the same time ,It has elaborated the method of Antarctic area map projection ,and has made use of the Universal Transverse Mercator Pojection (U TM ).K ey w ords :WalkField ;GIS Software ;South Pole ;Projection收稿日期:2007-05-08项目来源:国家测绘局南极重点项目作者简介:王连仲(1963～),男,高级工程师. 南极长城站工程地形图测绘是我国第23次南极科学考察队的重要任务之一,也是我国首次对南极地区进行大比例尺全野外数字化地形图测绘。

米据生，男，1966年3月生，博士，教授，博士生导师，河北师范大学学术带头人。

1986年毕业于河北师范大学数学系，1992年毕业于华东师范大学数学系，获理学硕士学位。

于2000年至2003年在西安交通大学攻读博士学位并开始从事信息系统知识获取方面的研究工作。

博士毕业后在香港中文大学进行博士后研究，主要从事概念格和粒计算的理论及其在地理信息系统中的应用的研究工作。

其后，多次应邀赴香港中文大学合作研究。

现主持国家自然科学基金项目一项和河北省自然科学基金项目一项。

作为主要研究人员曾参加两项国家自然科学基金项目的研究工作。

现任中国人工智能学会粗糙集与软计算专业委员会委员，国际期刊主编，多个国内、国际学术会议程序委员会委员。

主要研究兴趣：粗糙集、概念格、近似推理。

近几年在国际学术期刊《Information Sciences》、《International Journal of General Systems》、《International Journal of Intelligent Systems》、《International Journal of Approximate Reasoning》、《Computer & Mathematics with Applications 》、《Fuzzy Set and Systems》，国内核心学术期刊如《计算机学报》、《模式识别与人工智能》等和国际重要学术会议发表学术论文40余篇，其中13篇被SCI收录，20余篇被EI收录现主持的基金项目：1. 基于概念格的属性约简与知识获取方法研究，国家自然科学基金（NO:60773174）负责人，2008-2010. 29万。

2. 概念格属性约简方法与知识获取，河北省自然科学基金（NO:A2006000129）负责人，2006.1-2008.12，3万。

主要论文成果：[1]Ju-Sheng Mi, Yee Leung, Hui-Yin Zhao, Tao Feng, Generalized fuzzy rough setsdetermined by a triangular norm, Information Sciences, 2008. SCI.[2]Ju-Sheng Mi, Yee Leung, and Wei-Zhi Wu, An uncertainty measure inpartition-based fuzzy rough sets, International Journal of General Systems 34, (2005): 77-90. SCI, EI[3]Ju-Sheng Mi, Wei-Zhi Wu, and Wen-Xiu Zhang, Approaches to knowledgereduction based on variable precision rough sets model, Information Sciences 159(3-4), (2004) 255-272. SCI, EI.[4]Ju-Sheng Mi, Wen-Xiu Zhang, An axiomatic characterization of a fuzzygeneralization of rough sets, Information Sciences 160(1-4) 235-249. 2004. SCI, EI.[5]Yee Leung, Manfred M. Fischer, Wei-Zhi Wu, Ju-Sheng Mi, A rough set approachfor the discovery of classification rules in interval-valued information systems, International Journal of Approximate Reasoning 47(2) 2008, SCI, EI[6]Mi Ju-Sheng, Li Xiu-Min, Zhao, Hui-Yin, Feng Tao, Information-theoreticmeasure of uncertainty in generalized fuzzy rough sets. LECTURE NOTES IN ARTIFICIAL INTELLIGENCE 4482: 63-70, 2007.ISTP, EI.[7]HUI-YIN ZHAO, JU-SHENG MI, THE UNIFIED LOWER AND UPPERAPPROXIMATIONS BASED ON A FUZZY IDEAL, Proceedings of the SixthInternational Conference on Machine Learning and Cybernetics, Hong Kong, 19-22 August 2007. 1226-1231. EI[8]Yee Leung; Tung Fung; Ju-Sheng Mi; Wei-Zhi Wu, A rough set approach to thediscovery of classification rules in spatial data, International Journal of Geographical Information Science, 21(9), 2007, 1033–1058. SCI, EI.[9]Mi, Ju-Sheng; Leung, Yee; Zhao, Hui-Yin, The minimal sets of axiomscharacterizing rough fuzzy approximation operators. Proceedings of the 2006 International Conference on Machine Learning and Cybernetics, 2006, August, 1138-1143. EI[10]W u Wei-Zhi, Mi Ju-Sheng, Knowledge reduction in incomplete informationsystems based on Dempster-Shafer theory of evidence, In: Wang G., Peters J. F., Skowron A., Yao Y. Y.(Eds.), The 1st International Conference on Rough Sets and Knowledge Technology, LECTURE NOTES IN ARTIFICIAL INTELLIGENCE 4062(2006): 254-261. EI, ISTP.[11]F eng Tao, Mi Ju-Sheng, Wu Wei-Zhi, Covering-based generalized rough fuzzysets, Lecture Notes in Artificial Intelligence 4062(2006): 208-215. ISSN 0302-9743；2006年9月; SCI, EI[12]W ei-Zhi Wu, Yee Leung, and Ju-Sheng Mi, On characterizations of (I, T)-fuzzyrough approximation operators, Fuzzy Sets and Systems 154 (1), (2005): 76-102.SCI, EI.[13]W ei-Zhi Wu, Mei Zhang, Huai-Zu Li, and Ju-Sheng Mi，Knowledge reduction inrandom information systems via Dempster-Shafer theory of evidence, Information Sciences 2005,174 (3-4): 143-164, SCI, EI.[14]J u-Sheng Mi, Zeng-Xiao Guo, Tao Feng, Hui-Yin Zhao, Uncertainty inGeneralized Fuzzy Rough Sets, Proceedings of IEEE INTERNATIONAL CONFERENCE ON GRANULAR COMPUTING, 213-216, 2005. JUL Y, Beijing China. EI, ISTP[15]W en-Xiu Zhang, and Ju-Sheng Mi, Incomplete information system and itsoptimal selections, Computers & Mathematics with Applications 48(5-6): 691-698 SEP 2004. SCI, EI.[16]J u-Sheng. Mi, Dega. Chen, Wen-Xiu Zhang, Stochastic Aumann Integral,Southeast Asian Bull. Math., (2004), 27, 851-860.[17]W EI-ZHI WU, JU-SHENG MI, HUAI-ZU LI，A GENERAL VERSION OFKNOWLEDGE REDUCTION IN R-INFORMATION SYSTEMS. Proceedings of the Third International Conference on Machine Learning and Cybernetics, Shanghai, 26-29 August 2004. EI.[18]米据生，张文修，吴伟志，基于变精度粗糙集理论的知识约简方法，系统工程理论与实践2004 (1).76-82. EI.[19]J u-Sheng Mi, Wei-Zhi Wu, and Wen-Xiu Zhang, Approaches to ApproximationReducts in Inconsistent Decision Tables. LECTURE NOTES IN ARTIFICIAL INTELLIGENCE 2639. SCI, EI.[20]W en-Xiu Zhang, Ju-Sheng Mi, and Wei-Zhi Wu, Approaches to knowledgereductions in inconsistent systems, International Journal of Intelligent Systems，18 (2003), 989-1000. SCI, EI.[21]W u Wei-Zhi，Mi Ju-Sheng, and Zhang Wen-Xiu , Generalized fuzzy rough sets,Information Sciences 151 (2003)263-282. SCI, EI.[22]M i Jusheng, Chen Degang, and Zhang Wenxiu, Fuzziness in Generalized FuzzyRough Sets, Information, an International Journal. 6(4)(2003): 365-372.[23]C hen Degang, Mi Ju-Sheng, Some notes on the supremum and infimum of thefuzzy number space. J. Fuzzy Math. 11 (2003), 811-816.[24]M i Ju-Sheng，Wu Wei-Zhi，and Zhang Wen-Xiu , Convergence theorems forset-valued conditional expectations, Chinese Journal of Applied Probability and Statistics （应用概率统计）（2003）19 (1) 60-64.[25]W EI-ZHI WU, JU-SHENG MI, WEN-XIU ZHANG, A NEW ROUGH SETPPROACH TO KNOWLEDGE DISCOVERY IN INCOMPLETE INFORMATION SYSTEMS, Proceedings of the Second International Conference on Machine Learning and Cybernetics, Xi’an, 2-5 November 2003, EI, ISTP [26]米据生，吴伟志，张文修，不协调目标信息系统知识约简的比较研究，模糊系统与数学，（2003）17（3）：54-60.[27]M i Ju-Sheng, Zhang Wen-Xiu, Wu Wei-Zhi, Optimal decision rules based oninclusion degree theory. IEEE of the First International Conference on Machine Learning and Cybernetics, Beijing, 2002, 1223--1226. EI, ISTP [28]米据生, 吴伟志，张文修（2002），粗糙集的构造与公理化方法. 模式识别与人工智能，15（3）: 280-284. EI[29]米据生，张文修，基于粗糙集的间接学习，计算机科学，29（6）(2002).96-97,104.[30]W u Wei-Zhi, Mi Ju-Sheng, and Zhang Wen-Xiu (2002), Combination ofapproximation spaces and its applications，工程数学学报, 19(3), 86-94.[31]J usheng Mi, Wenxiu Zhang, Composition of General Fuzzy ApproximationSpaces, Lecture Notes in Artificial Intelligence 2275, 2002, 497-501教材：[1]杨林生,乔玉英，米据生，复变函数，高等教育出版社, 2000.[2]米据生，副主编，经济应用数学，地质大学出版社, 1997.[3]米据生，副主编，经济数学（上、下），吉林大学出版社, 2000.另附照片一张。