智慧財產權對於現今企業是一道很重要的課題，而智慧财產權包含有專利、商標及著作權等等，這些都是企業競爭發展過程中很重要的商業武器，如何獲得這些商業武器，均有一個很重要的動作，也是企業發展過程常會遺忘的動作---“申請”，因為世界各國之專利、商標係採“先申請主義”，此為其一注意之處，其二就是不論是專利或商標在申請前的檢索動作，此對於專利或商標案件申請後之後續核准有著決定性的影響，但鑒於專利要件判斷更有其主觀性，很容易受到於各別審查委員的心證影響而有所區別外，同時專利更是一份技術文獻與權利界定書，其判別上更有其特殊性、專業性，鑒於可規範的部份比較複雜，所以不在本研究的考慮範圍，故本研究主要針對商標申請前之圖形檢索部份加以探討與精進模式。
商標的組成不外是文字或圖形等其一單獨存在，或者為文字與圖形相互組合而成等形態，鑒於單純的文字判別上只要針對文字在資料庫中搜尋，即可快速知道有無相同之文字申請在先，因此檢索上相對容易且依現有文字比對在資料庫系統並無困難之處，因此沒有改善之急迫性，反而在商標圖形部份，卻因為商標圖形設計因素，容易呈現出非固定型態，再加上現有資料庫檢索之種種限制，不論在時間上或人力耗費上，均相當龐大，鑒於現有商標資料庫之檢索，對於欲註冊之商標圖形後續之核准判斷影響甚深，期能提出一種利用機器人流程自動化(RPA)搭配AI圖形辨識等創新科技輔助之檢索模式，再輔以核駁案件之審查要件的教示、提點，藉此綜合判斷得以改善現有商標圖形檢索模式之精準度，進而減少商標圖形申請中之不確定性。

Nowadays, the intellectual property right (IPR) is a very important issue to enterprises. The most common intellectual property right includes patents, trademarks, copyrights, etc. These are very important business weapons in the competition among enterprises. “Filing” is of critical for getting these businesses weapons, which is, however, easily ignored during the development of the enterprises. Two points should be considered. Firstly, patent and trademark applications around the world are examined on the basis of a “First-to-File” principle. Secondly, a prior art search conducted before filing patent and trademark applications greatly affects the subsequent development of applications. However, the patentability for patent applications is determined according to subjective judgments of different examiners, which easily causes different examination results. In addition, a patent application involves technical contents and claims, and therefore the judgment inevitably requires particularities and professionalism. Because the regulation applicable to the patent examination is special and more complicated, it is not considered in the scope of this study. This study mainly focuses on the discussion about a figurative mark search prior to the filing of trademark applications and its corresponding improvement.
A trademark application usually consists of word(s), figurative elements or symbols, or any combination of word(s) and figurative elements or symbols. When a trademark application is a word mark simply typed in words or letters, a prior word search can be quickly conducted in the search database to check whether there are earlier-filed word marks. In other words, it is easier to conduct a prior word mark search by comparing the stored words in the search database with words or letters typed in the trademark application. Hence, the improvement in the word search is not an emergency. In contrast, a figurative trademark application easily involves unfixed models because of various figurative, graphic or symbol designs, and the prior figurative mark search in the present search database is limited by many factors. For instance, a figurative mark search might be time-consuming or costly. Since the trademark search in the present search database is of critical in determining whether a trademark right shall be granted, it is desirable to create an innovative search model by the cooperation between the robotic process automation (RPA) and the AI pattern recognition, with the aid of teaching and advice as suggested by criteria against the applications for examinations, and the innovative search model thereby attains a synthetic judgment whereby the searching accuracy in the present search model for the figurative mark search is improved to reduce uncertainties for the filing of the figurative trademark application.

論文審定書 i
致謝 ii
摘要 iii
Abstract iv
目 錄 vi
圖 次 ix
表 次 xi
第一章 緒論 1
第一節 研究背景 1
第二節 研究動機 4
第三節 研究問題 9
一、現行商標圖形申請之比對系統 9
二、智慧財產局之商標圖形檢索系統之流程與問題 11
第四節 研究目的 13
第二章 文獻探討 14
第一節 商標及其申請審查流程 14
一、商標法規即其定義 14
二、商標分類 16
三、商標審查基準及其流程 17
四、商標相似案例 18
第二節 設計科學法 22
第三節 創新科技之運用 23
一、前言 23
二、分析 24
三、下載資料內容 27
第四節 AI人工智慧之圖形辨識 28
一、前言 28
二、卷積與池化 29
三、完全連接層 32
第三章 設計科學法 35
第一節 前言 35
第二節 設計科學法流程 35
第三節 研究方法與步驟 37
一、確認問題與動機 37
二、解決方案與目標的定義 38
三、設計與發展 39
四、展示 39
五、評估與溝通 39
第四章 系統設計與建置 40
第一節 系統架構 40
第二節 資料處理 41
一、圖樣比對因素 41
二、本研究規劃檢索系統 41
第五章 研究內容與發現 42
第一節 前言 42
第二節 訪問 42
第三節 訪問內容摘要 43
第四節 統整問卷 45
第六章 結論與建議 46
第一節 研究與結論 46
第二節 研究建議 47
第三節 研究限制 47
第四節 未來方向 48
參考文獻 49
附錄一 53
附錄二 60
附錄三 65

【1】M. F. Omar, B. Trigunarsyah and J. Wong, "A Design Science Approach for Consultant Selection Decision Support System," 2009 Fourth International Conference on Cooperation and Promotion of Information Resources in Science and Technology, 2009, pp. 90-94, doi: 10.1109/COINFO.2009.73.

【2】L. Abdurrahman and T. Mulyana, "Parallel Construction of Information Technology Value Model: Design-Science Research Methodology," 2020 8th International Conference on Information and Communication Technology (ICoICT), 2020, pp. 1-8, doi: 10.1109/ICoICT49345.2020.9166214.

【3】http://www.naipo.com/Portals/1/web_tw/Knowledge_Center/Industry_Economy/IPNC_200527_0701.htm

【4】D. Keysers, T. Deselaers, C. Gollan and H. Ney, "Deformation Models for Image Recognition," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, no. 8, pp. 1422-1435, Aug. 2007, doi: 10.1109/TPAMI.2007.1153.

【5】https://demo.leemeng.tw/

【6】https://ithelp.ithome.com.tw/articles/10191820

【7】Wiki , "Neuron Hand-tuned" , https://zh.wikipedia.org/wiki/File:Neuron_Hand-tuned.svg。

【8】Y. Ketkar and S. Gawade, "Effectiveness of Robotic Process Automation for data mining using UiPath," 2021 International Conference on Artificial Intelligence and Smart Systems (ICAIS), 2021, pp. 864-867, doi: 10.1109/ICAIS50930.2021.9396024.

【9】CHRISTOPHER M. BISHOP ," Neural Networks for Pattern Recognition ".

【10】S. Sutipitakwong and P. Jamsri, "The Effectiveness of RPA in Fine-tuning Tedious Tasks," 2020 6th International Conference on Engineering, Applied Sciences and Technology (ICEAST), 2020, pp. 1-4, doi: 10.1109/ICEAST50382.2020.9165452.

【11】R. Uskenbayeva, Z. Kalpeyeva, R. Satybaldiyeva, A. Moldagulova and A. Kassymova, "Applying of RPA in Administrative Processes of Public Administration," 2019 IEEE 21st Conference on Business Informatics (CBI), 2019, pp. 9-12, doi: 10.1109/CBI.2019.10089.

【12】臺北高等行政法院 95 年度訴字第 404 號判決, https://law.judicial.gov.tw/FJUD/data.aspx?ty=JD&id=TPBA,95%2c%e8%a8%b4%2c404%2c20070131%2c2

【13】臺北高等行政法院 96 年度訴字第 1374 號判決, https://law.judicial.gov.tw/FJUD/data.aspx?ty=JD&id=TPBA,96%2c%e8%a8%b4%2c1374%2c20071213%2c2

【14】臺灣臺北地方法院 102 年度智簡字第 30 號刑事判決,
https://law.judicial.gov.tw/FJUD/data.aspx?ty=JD&id=TPDM,102%2c%e6%99%ba%e7%b0%a1%2c30%2c20130329%2c1

【15】商標註冊常見問題, http://www.cantoop.com/news/newsdetail_25.shtml

【16】經濟部智慧財產局商標業務統計,
https://topic.tipo.gov.tw/trademarks-tw/lp-583-201.html

【17】S. Yatskiv, I. Voytyuk, N. Yatskiv, O. Kushnir, Y. Trufanova and V. Panasyuk, "Improved Method of Software Automation Testing Based on the Robotic Process Automation Technology," 2019 9th International Conference on Advanced Computer Information Technologies (ACIT), 2019, pp. 293-296, doi: 10.1109/ACITT.2019.8780038.

【18】S. Parchande, A. Shahane and M. Dhore, "Contractual Employee Management System Using Machine Learning and Robotic Process Automation," 2019 5th International Conference On Computing, Communication, Control And Automation
(ICCUBEA), 2019, pp. 1-5, doi: 10.1109/ICCUBEA47591.2019.9128818.
【19】L. Lu, Y. Yi, F. Huang, K. Wang and Q. Wang, "Integrating Local CNN and Global CNN for Script Identification in Natural Scene Images," in IEEE Access, vol. 7, pp. 52669-52679, 2019, doi: 10.1109/ACCESS.2019.2911964.

【20】X. -H. Han, Y. Sun and Y. -W. Chen, "Residual Component Estimating CNN for Image Super-Resolution," 2019 IEEE Fifth International Conference on Multimedia Big Data (BigMM), 2019, pp. 443-447, doi: 10.1109/BigMM.2019.00028.

【21】D. Keysers, T. Deselaers, C. Gollan and H. Ney, "Deformation Models for Image Recognition," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, no. 8, pp. 1422-1435, Aug. 2007, doi: 10.1109/TPAMI.2007.1153.

【22】G. Showkatramani, N. Khatri, A. Landicho and D. Layog, "Deep Learning Approach to Trademark International Class Identification," 2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA), 2019, pp. 608-612, doi: 10.1109/ICMLA.2019.00112.

【22】Wing Ho Leung and Tsuhan Chen, "Trademark retrieval using contour-skeleton stroke classification," Proceedings. IEEE International Conference on Multimedia and Expo, 2002, pp. 517-520 vol.2, doi: 10.1109/ICME.2002.1035662.

【23】Alwis and Austin, "An integrated framework for trademark image retrieval using gestalt features and CMM neural network," Image Processing And Its Applications, 1999. Seventh International Conference on (Conf. Publ. No. 465), 1999, pp. 290-295 vol.1, doi: 10.1049/cp:19990329.

【24】F. Schorr and L. Hvam, "Design Science Research: A Suitable Approach to Scope and Research IT Service Catalogs," 2018 IEEE World Congress on Services (SERVICES), 2018, pp. 25-26, doi: 10.1109/SERVICES.2018.00026.

【25】智慧財產局商標檢索系統https://twtmsearch.tipo.gov.tw/OS0/OS0302.jsp?l6=zh_TW&isReadBulletinen_US=&isReadBulletinzh_TW=true

【26】卷積神經網路, https://m.xuite.net/blog/metafun/life/589355242

【27】商標註冊流程, https://ezstartup.cc/apply-trademark.html

【28】[資料分析&機器學習] 第5.1講: 卷積神經網絡介紹
https://medium.com/jameslearningnote/%E8%B3%87%E6%96%99%E5%88%86%E6%9E%90-%E6%A9%9F%E5%99%A8%E5%AD%B8%E7%BF%92-%E7%AC%AC5-1%E8%AC%9B-%E5%8D%B7%E7%A9%8D%E7%A5%9E%E7%B6%93%E7%B6%B2%E7%B5%A1%E4%BB%8B%E7%B4%B9-convolutional-neural-network-4f8249d65d4f

【28】解析設計研究法的架構與實施： 以科學教育研究為例 http://rportal.lib.ntnu.edu.tw:8080/server/api/core/bitstreams/2fb09588-ec81-4add-a3c2-be24b8035829/content

【29】商標法https://psla.nkuht.edu.tw/var/file/15/1015/img/37/227855501.pdf

【30】商標法制之建構～從歷次商標法修正觀察
https://chien-yeh.com.tw/tw/2019/10/01/%E5%95%86%E6%A8%99%E6%B3%95%E5%88%B6%E4%B9%8B%E5%BB%BA%E6%A7%8B%EF%BD%9E%E5%BE%9E%E6%AD%B7%E6%AC%A1%E5%95%86%E6%A8%99%E6%B3%95%E4%BF%AE%E6%AD%A3%E8%A7%80%E5%AF%9F/

【31】應用設計科學研究法(DSRM)建置商業智慧系統之研究-以科學園區物流運輸服務業為例,https://hdl.handle.net/11296/8k5xd8

【32】基於以太坊技術之交易代幣化設計科學研究,https://hdl.handle.net/11296/432ug4

【33】以形狀為基礎之葉子影像檢索系統，中華大學碩士論文，作者:許志祥，民國94年。http://chur.chu.edu.tw/bitstream/987654321/3298/1/NC093CHPI0392067.pdf