臺灣博碩士論文加值系統

問卷調查為健康醫療領域常使用的資料蒐集方法，其實施上不僅耗費人力與金錢，也常因難以避免的人為輸入錯誤而造成困擾。本研究目的是希望能夠透過設計出一個自動化問卷辨識系統來支援整個問卷調查的生命週期，並且希望可以利用其來提高研究調查的品質，加速問卷調查的流程，與降低問卷調查的成本。

本研究主要利用光學標誌辨識(OMR)技術，來研究與設計出自動化問卷辨識系統。使用者將可使用其來取代傳統的人工問卷資料輸入，其將可迅速且有效地辨識出問卷結果，並可彈性地搭配文書處理軟體(Microsoft Office Word)來做相關的問卷設計工作。在系統架構上主要可分為三個主要的模組，分別是文件定位產生器(DPG: Document Positioning Generator)、問卷辨識引擎(QRE :Questionnaire Recognition Engine)、結果校正與統計工具(CST :Correction and Statistic Tools)，而為了提高資料轉移的可行性，系統中資料儲存方式採用XML的資料交換格式，如此將來我們也可以輕易地將資料轉移到其他的資料庫系統中。

在系統操作流程方面，第一步須先完成問卷設計與利用DPG來建立該問卷的譯碼簿(code book)，然後便可接著利用DPG來將紙本問卷原稿之電子檔(.doc)或掃瞄圖檔(.bmp)做事先的問卷定位工作。而之後回收的調查問卷可先透過附有自動快速送紙器(ADF)的掃瞄器來掃瞄成圖檔，也可利用傳真數據機來將問卷透過傳真的方式傳回成圖檔到電腦上。接著便可利用QRE來做問卷辨識的工作，然後便可利用CST來觀看辨識結果列表，並可直接透過圖形介面操作方式來直接做辨識結果校正的工作。此外最後也可以將問卷調查之統計結果匯入至其他之相關統計分析軟體(如: Microsoft® Office Excel或 SPSS : Statistical Package for the Social Sciences)來做進階的統計分析處理工作。

根據本系統的評估結果顯示在問卷原稿事先定位方面，直接利用其影像定位會較利用其電子檔(.doc)定位來的好；而在問卷辨識率方面，透過掃瞄器得到的問卷影像會較透過傳真得到的問卷影像來的好。目前系統的checkbox辨識錯誤率最低可以達到0.00057，而每頁問卷所需的處理時間(包含問卷影像擷取與辨識)最少只需1.17秒。

本系統將被定位為一個介於傳統紙本問卷調查與現今電子化問卷調查之間的中介系統，而因為較低人力成本與人為錯誤率，其將可用於取代或協助傳統的人工閱卷方式。系統功能上目前仍有不足的地方，而未來我們將仍會繼續地努力去改善。

Questionnaires are generally used for collecting data in public health and medicine. But they always consume much manpower and money, and the problem of human errors also occurs frequently when entering data. The purpose of our research is to design an automatic questionnaire recognition system to support entire life cycle of survey. And we hope this system could help to improve the quality of surveys, speed up the flow of surveys and reduce the cost of surveys.

The main technique we used to design this system was OMR (Optical Mark Recognition). The system included three primary modules which were “Document Positioning Generator (DPG)”, “Questionnaire Recognition Engine (QRE)” and “Correction and Statistic Tools (CST)”. The output of our system is in XML-format which makes the data easily transferable to other system and format.

The first step of system operating flow was to finish the questionnaire design and to use the DPG to build the code book of questionnaire, and then users could use the DPG to do the pre-position work by electronic file (.doc) or scanning image (.bmp) of questionnaire draft. Afterward the answered questionnaire images could be collected by scanning with scanner has ADF (automatic document feeder) or directly faxing to image files to computer through fax modem. The next step was to use the QRE to start questionnaire recognition work, and then users could use the CST to view the recognition results and statistical table. Besides, users could also use the CST to directly correct the recognition results by GUI (Graphical User Interface) tool. Finally, the statistical result could also be transformed into a text file. The text file which could be used by other statistical software to do advanced statistical analysis and produce the related statistical tables or charts, e.g. Microsoft® Office Excel and SPSS (Statistical Package for the Social Sciences).

According to the evaluation result of our system, directly using the paper-image of questionnaire draft to pre-position is better than using electronic file (.doc) of questionnaire draft. And using the questionnaire images from scanner to recognize is better than questionnaire images form fax. The lowest error rate of questionnaire recognition is 0.00057 per checkbox, and the lowest time of questionnaire recognition which include gathering and recognition of questionnaire images is 1.17 seconds per page.

This system will be a junction system between traditional paper-based survey and current electronic survey. And because of the lower human errors and the lower manpower users could use this system to replace the traditional manual data entry. But this system is still insufficient for user’s needs. We will endeavor to improve our system continually.

[1] Laura Landro. Push Grows for Online Health Data. http://online.wsj.com/ . 3-11-2004. The Wall Street Journal Online.
[2] Barbara B.Moran. Virtual Realists: Librarians in a Time of Transition. North Carolina Libraries 1999; 57(4):165-169.
[3] Mark Hagland. Document Management Systems. http://www.healthcare-informatics.com/issues/2005/04_05/snapshot.htm . 2005. Healthcare Informatics Online.
[4] Stephen R.Jones. Keynote: Dr. Charles Geschke at PC Expo 99 New York. http://www.reviewsonline.com/PCX99KY3.HTM . 6-24-1999.
[5] James T.Mulder. Hospitals Going Paperless. http://www.syracuse.com/business/poststandard/index.ssf?/base/business-6/111062036127250.xml . 3-13-2005.
[6] Pincus T. Documenting quality management in rheumatic disease: are patient questionnaires the best (and only) method? Arthritis Care & Research 1996; 9:339-348.
[7] Slack WV, Slack CW. Patient-Computer dialog. N Engl J Med 1972; 286:1304-1309.
[8] Eric Rose. Life After Go-Live-Part 1: Paper in the Paperless Practice. Journal of Healthcare Information Management 2003; 17(1):24-26.
[9] Hodsdon D.F. Remarks from a representation of farm workers, Ergonomics problems in agriculture. 1967. Meeting at Nottingham University School of Agriculture.
[10] Davis FD. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology. MIS Quarterly 1989; 13(3):319-340.
[11] Michael G.Morris, Andrew Dillon. The Influence of User Perceptions on Software Utilization: Application and Evaluation of a Theoretical Model of Technology Acceptance. IEEE Software 1997; 14(4):58-76.
[12] Michael E.Wiklund. Medical Device and Equipment Design-Usability Engineering and Ergonomics. Interpharm Press, Inc., 1995.
[13] Lu Ann Aday. Designing and Conducting Health Surveys: A Comprehensive Guide. San Francisco, California: Jossey-Bass Inc., 1989.
[14] OMR. http://www.le.ac.uk/TALENT/book/glossary.htm . 12-20-1999. Book of TALENT Glossary.
[15] Glossary of Terms-OMR. http://www.ballyclarehigh.co.uk/garden91/Glossary_finalRW.htm . 2005.
[16] OMR. http://en.wikipedia.org/wiki/OMR . 2-28-2005. Wikipedia.
[17] James J.Neutens, Laurna Rubinson. Research Techniques for the Health Sciences. 3 ed. San Francisco, CA.: Benjamin Cummings, Inc., 2002.
[18] Czaja S. Special issue preface. Human Factors 1990; 32(5):505.
[19] Diffrient N, A.Tilley, J.Bardagjy. Humanscale 4/5/6. Cambridge, MA: MIT Press. In press.
[20] LeGros Clark W.E. The anatomy of work. Floyd W.F., Welford A.T., editors. 1954. Lewis, London., Symposium on human factors in equipment design.
[21] MB Weigner, CE Englund. Ergonomic and human factors affecting anesthetic vigilance and monitoring performance in the operating room environment. Anesthesiology 1990; 73(5):995-1021.
[22] Anonymous. The next level of Internet applications: providing quality care to high-risk patients. Qual Lett Healthc Leaders 2000; 12(5):2-3.
[23] IT Discovery: Web Survey Project Final Report. http://web.mit.edu/is/discovery/web-surveys/report/ . 9-5-2001. Massachusetts Institute of Technology.
[24] Scott E Umbaugh. Computer vision and Image Processing: A Practical Approach Using CVIPtools. Prentice Hall PTR, 1998.
[25] Sing-Tze Bow. Pattern Recognition and Image Preprocessing. Marcel dekker, Inc., 1992.
[26] Neil Collings. Optical Pattern Recognition Using Holographic Techniques. Addison-Wesley Publishing Company, 1988.
[27] PC OMR V6.0. http://www.cai2000.com/index.html . 2005. SiQiSoft Company.
[28] LEADTOOLS Optical Mark Recognition Tools. http://www.leadtools.com/SDK/Document/Document-Addon-OMR.htm . 2005. LEAD Technologies, Inc.
[29] LEADTOOLS Document Imaging Suite SDK. http://www.leadtools.com/SDK/Document/Document-Imaging-Suite.htm . 2005. LEAD Technologies, Inc.
[30] LEADTOOLS OCR Programming Tools. http://www.leadtools.com/SDK/Document/Document-Addon-OCR.htm . 2005. LEAD Technologies, Inc.
[31] Paul Cornell. Working with the Office XP Primary Interop Assemblies. http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnoxpta/html/odc_oxppias.asp . 2002. Microsoft MSDN.
[32] Linda G.Shapiro, George C.Stockman. Computer Vision. Prentice Hall, Inc., 2001.