在人機互動介面演變過程中，現今的電腦多以圖形使用者介面的方式呈現，與早期電腦使用的命令列介面相比，在視覺上更為簡單、直覺，操作使用也更為方便。

在圖形介面中，圖示(Icons)為用來指示使用者執行各種操作的圖像，其中文件類型的檔案常以特定圖示表達，正因如此，造成文件之間在視覺上因而缺少了獨特性，使用者也無法從圖示中迅速喚起對此文件之內容的記憶，因而，搜尋文件時也會增加其搜尋時間。因此本論文提出了一種能自動生成具有文件上下文語義的圖示系統，透過文件分析找出關鍵字並且檢索出其關鍵字之對應圖像，運用視覺顯著區域偵測與圖像分割來提取圖像前景，最後以拼貼方式產生出代表此文件的圖示，以取代傳統圖形使用者介面中的圖示，提高檔案之間的獨特性與降低搜尋時間。

除了文件圖示研究之外，本論文也提出其他文件呈現之研究，包括利用文字與情緒分析將文件以視覺與聽覺的方式呈現給使用者。最終透過本研究之系統在尚未開啟檔案前以不同感受呈現文件，讓使用者對其文件產生更多記憶點，進而快速喚起對文件內容的記憶。

In the evolution of human-computer interface, modern desktop applications make use of the prevalent graphical user interfaces.It’s visually simpler, intuitive, and easier for users to work with.

Icons are parts of the graphical user interfaces for a computer system. They are pictograms displayed on a computer screen in order to help a user navigate a computer system. In general, the same type of documents are often represented by the same icon, so it may lead to lots of identical icons that are not distinctive. As users can’t quickly recall the contents of the document from the icons so it will increase the time of retrieval. Our research provides a system which can automatically generate icons associated with a document’s context semantics. Firstly we find keywords through document contents analysis and retrieve the images of keywords. Next, we make use of saliency detection and image segmentation to extract the image foreground. Finally, the icon is generated from a image collage. The resulting icons can enhance the representation of files in a Graphical User Interface by providing semantically and graphically distinguishable symbols.

In addition to icons, we also provide other presentations for documents. Using text analysis and sentiment analysis we present documents by auditory and visual information. Through our system, users can improve their associated memory and quickly evoke the contents of the documents by different feelings before the file is opened.

[1] Beatpick. http://www.beatpick.com/player/index.
[2] Incompetech. https://incompetech.com/music/royalty-free/music.html.
[3] Jieba. https://github.com/anderscui/jieba.NET.
[4] Pixabay. https://pixabay.com.
[5] H. B. Barlow. Possible principles underlying the transformations of sensory messages. IEEE Sensory Communication, page 18, 1961.
[6] Andreas Girgensohn, Frank M. Shipman III, Francine Chen, and Lynn Wilcox. Docubrowse: faceted searching, browsing, and recommendations in an enterprise context. In Proceedings of the 15th International Conference on Intelligent User Interfaces, IUI 2010, Hong Kong, China, February 7-10, 2010, pages 189–198, 2010.
[7] Xiaodi Hou and Liqing Zhang. Saliency detection: A spectral residual approach. In 2007 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007), 18-23 June 2007, Minneapolis, Minnesota, USA, 2007.
[8] Laurent Itti, Christof Koch, and Ernst Niebur. A model of saliency-based visual attention for rapid scene analysis. IEEE Trans. Pattern Anal. Mach. Intell.,20(11):1254–1259, 1998.
[9] Philipp Kolhoff, Jacqueline Preuß, and J¨orn Loviscach. Music icons: Procedural glyphs for audio files. In 19th Brazilian Symposium on Computer Graphics and Image Processing (SIBGRAPI 2006), 8-11 October 2006, Manaus, Amazonas,Brazil, pages 289–296, 2006.
[10] John P. Lewis, Ruth Rosenholtz, Nickson Fong, and Ulrich Neumann. Visualids: automatic distinctive icons for desktop interfaces. ACM Trans. Graph.,23(3):416–423, 2004.46
[11] Nobuyuki Otsu. A threshold selection method from gray-level histograms. IEEE Transactions on Systems, Man and Cybernetics, page 5, 1979.
[12] Carsten Rother, Vladimir Kolmogorov, and Andrew Blake. ”grabcut”: interactive foreground extraction using iterated graph cuts. ACM Trans. Graph.,23(3):309–314, 2004.
[13] Rishiraj Saha Roy, Abhijeet Singh, Prashant Chawla, Shubham Saxena, and Atanu R. Sinha. Automatic assignment of topical icons to documents for faster file navigation. In 14th IAPR International Conference on Document Analysis and Recognition, ICDAR 2017, Kyoto, Japan, November 9-15, 2017, pages 1338–1345, 2017.
[14] Vidya Setlur, Conrad Albrecht-Buehler, Amy Ashurst Gooch, Samuel Rossoff, and Bruce Gooch. Semanticons: Visual metaphors as file icons. Comput. Graph.Forum, 24(3):647–656, 2005.
[15] Hendrik Strobelt, Daniela Oelke, Christian Rohrdantz, Andreas Stoffel, Daniel A. Keim, and Oliver Deussen. Document cards: A top trumps visualization for documents. IEEE Trans. Vis. Comput. Graph., 15(6):1145–1152,2009.
[16] Min-Joon Yoo and In-Kwon Lee. Affecticon: Emotion-based icons for music retrieval. IEEE Computer Graphics and Applications, 31(3):89–95, 2011.