因智慧型手持裝置與網際網路的普及應用，使得生活型態與之劇變。智慧型手持裝置像台迷你電腦，除了通話功能外大幅提升生活便利性。原本許多需在特定地方及物品(如：桌上型電腦)的功能也在其中，也使得使用者對智慧型手持裝置的粘著度越來越高。許多行動應用也以「無所不在」的服務為起點。
本研究是利用手持裝置的相機功能搭配網際網路或電腦搭配網際網路，將圖像或PDF文件上傳至Web經由OCR文字辨識系統即時進行辨識，將資訊回傳至手機，取代打字形式來獲得訊息，再經由NLP自然語言分析擷取文件內容資訊的關鍵字或關鍵詞來輔助使用者閱讀，若使用者對於關鍵字或詞有不了解的地方，只需點擊其字詞或將不了解的詞彙反白即可馬上由維基百科查尋詞彙解釋。希望透過本研究的行動應用，可協助使用者在閱讀及知識累積上更多的便利性。

Lifestyles are changing with the spread of smart handheld devices and development of the Internet. A smart handheld device is like a mini computer, and it can greatly improve the convenience of life, in addition to be used as a phone. Many of the features that were originally needed in specific places and are associated with some devices (such as desktop computers) are also included, making smart handheld devices more and more important to users. Many mobile applications also start to offer "ubiquitous" services.
This study uses the camera function of a handheld device with the Internet correction or a computer with Internet connection to upload images or PDF files to the Web for instant identification through OCR capability, and send information back to mobile phones to obtain information. It is through NLP analysis to retrieve the keywords or keywords of the document content information to assist users in reading. If users do not understand the keywords or words, they can just click on the words they do not understand to immediately search for the word interpretation from Wikipedia. It is hoped that the application of this research can help users to make reading and knowledge accumulation much easier.

[1] "ADO." https://www.itread01.com/p/1407878.html (accessed June, 1, 2021).
[2] "ASP.NET Core." https://www.itread01.com/content/1558077002.html (accessed June, 1, 2021).
[3] "AVX-512." https://blog.51cto.com/u_15333820/3465327 (accessed June, 1, 2021).
[4] "Double Array Trie." https://www.itread01.com/content/1550088374.html (accessed June, 1, 2021).
[5] "Fast Float." https://www.itread01.com/content/1550634858.html (accessed June, 1, 2021).
[6] "hOCR." https://github.com/kba/hocr-spec/blob/master/1.1/spec_zh_CN.md (accessed June, 1, 2021).
[7] "Jieba." https://coderwall.com/p/38wtgw/jieba (accessed June, 1, 2021).
[8] "OCR." https://zh.wikipedia.org/wiki/光學字元辨識 (accessed June, 1, 2021).
[9] "OpenCV." https://zh.wikipedia.org/wiki/OpenCV (accessed June, 1, 2021).
[10] "Ostu." https://dotblogs.com.tw/dragon229/2012/12/28/86092 (accessed June, 1, 2021).
[11] "Python." https://zh.wikipedia.org/wiki/Python (accessed June, 1, 2021).
[12] "socket." https://www.itsfun.com.tw/Socket/wiki-832207 (accessed June, 1, 2021).
[13] "Tesseract 5.0 Beta." https://github.com/tesseract-ocr/tesseract/releases/tag/5.0.0-beta-20210815 (accessed June, 1, 2021).
[14] "TSV." https://www.reviversoft.com/zh-tw/file-extensions/tsv (accessed June, 1, 2021).
[15] "Understanding LSTM Networks." https://colah.github.io/posts/2015-08-Understanding-LSTMs/ (accessed June, 1, 2021).
[16] "長短期記憶." https://zh.wikipedia.org/wiki/長短期記憶 (accessed June, 1, 2021).
[17] "動態規劃." http://web.ntnu.edu.tw/~algo/DynamicProgramming.html (accessed June, 1, 2021).
[18] "循環神經網路." https://www.easyatm.com.tw/wiki/循環神經網路 (accessed June, 1, 2021).
[19] "隱馬可夫模型." https://www.cnblogs.com/skyme/p/4651331.html (accessed June, 1, 2021).
[20] "斷開中文的鎖鍊！自然語言處理." 中央研究院. https://newsletter.sinica.edu.tw/13928/ (accessed June, 1, 2021).
[21] 王昊奮、邵浩、李方圓、張凱、宋亞楠, 中文自然語言處理實戰. 博碩文化股份有限公司 (in 繁體中文), 2019.
[22] 林彥佑, "利用光學字符辨識技術所設計之手機版藥袋用藥資訊辨識系統," 碩士學位論文, 國立臺灣科技大學, 電機工程系, 2015.
[23] 翁瑋鴻, "基於WEB OCR引擎之即時文字資訊辨識行動應用程式," 碩士論文, 國立高雄第一科技大學, 電腦與通訊工程系, 2013.
[24] 楊恩承, "基於光學字元辨識技術的個人健康資訊系統於智慧型手機應用," 碩士論文, 中華大學, 電機工程學系, 2014.
[25] 董付國, Python也可以這樣學. 博碩文化股份有限公司(in 繁體中文), 2017.
[26] 賴昀偉, "基於影像處理與光學字元辨識之自動定位系統開發研究," 碩士學位論文, 國立臺灣科技大學, 自動化及控制研究所, 2021.