智慧型手機之功能性照片自動集合系統

SnapGroup: Supporting Grouping of Functional Photos Taken with Smartphones

王珣沛(Hsun-Pei Wang)

指導教授 ： 陳炳宇

國立臺灣大學/管理學院/資訊管理學研究所/碩士(2013年)

https://doi.org/10.6342/NTU.2013.01267

摘要

因為智慧型手機輕便、易於隨身攜帶的特性，在任何時刻隨心所欲的拍攝照片成為可能，也因為智慧型手機輕便隨身的關係，智慧型手機拍攝的照片與傳統數位相機所拍攝的有不同的特性。然而，現有的照片管理工具仍以類似於個人電腦中的設計來協助使用者進行照片的管理，我們從使用者訪談中我們發現，使用者對於現有的設計感到挫折，且不同用途的照片混雜在一起。
從使用者調查中，我們發現智慧型手機的照片可以被歸納成三大類別：分別是功能性照片（Functional Photos）、事件類型照片（Event Photos）以及生活隨拍（Random Snapshots）。支援這三類型的照片的整理，可以更方便使用者根據照片的特性進行搜尋與整理。由於如何協助快速整理功能性照片尚未被先前研究充分探索，我們將重點放在功能性照片的自動分類。
我們從14位使用者收集到個人以手機拍攝的功能性及非功能性的照片，透過我們結合人臉、紋理及顏色特徵的方式，能夠使ROC曲線下面積達到（AUC）0.861，能夠有效的分類出功能性手機照片。

關鍵字

照片集合 ； 自動照片分類 ； 手機照片 ； 使用者導向設計

並列摘要

With the portable nature and compactness of smartphones, users nowadays are now able to take photos of any moments they like, thus bringing about different behaviors of photography practices than conventional digital cameras. Existing photo organizational tools on smartphones and related literature inherit similar design used in personal computers. However, in our formative user study, most users felt frustrated organizing their photos taken with smartphones, and photos taken for different purposes are mixed together by current design.
We discovered from the user study that photos taken with smartphones can be summarized into three different categories - functional photos, event photos, and random snapshots. Supporting grouping of the three types of photos easily enables users to search and organize them more easily. Since supporting grouping of functional photos has not been well-explored, we put focus on discussing classifying functional photos automatically in this research.
We collected both functional photos and non-functional ones from 14 participants. By using our methods combining the face model with texture and color features, it is able to achieve AUC about 0.861, an encouraging result considering the complex semantics of photos.

並列關鍵字

Photo Grouping ； Automatic Photo Classification ； Smartphone Photos ； User-Centered Design

參考文獻

[1] Evernote. http://evernote.com/.

[5] M. Ames, D. Eckles, M. Naaman, M. Spasojevic, and N. House. Requirements for mobile photoware. Personal Ubiquitous Comput., 14(2):95–109, Feb. 2010.

[6] D. Borth, C. Schulze, A. Ulges, and T. M. Breuel. Navidgator - similarity based browsing for image and video databases. In Proc. KI ’08, pages 22–29, 2008.

[8] S. Chatzichristofis and Y. Boutalis. Fcth: Fuzzy color and texture histogram - a low level feature for accurate image retrieval. In Image Analysis for Multimedia Interactive Services, 2008. WIAMIS ’08. Ninth International Workshop on, pages 191–196, 2008.

[9] S. A. Chatzichristofis and Y. S. Boutalis. Cedd: color and edge directivity descriptor: a compact descriptor for image indexing and retrieval. In Proceedings of the 6th international conference on Computer vision systems, ICVS’08, pages 312–322, Berlin, Heidelberg, 2008.