手機是近代以來最為普遍的消費性電子產品之一。而智慧型手機除了具有傳統手機所具備的通話簡訊等基本功能之外，使用者還可以利用其開放式的作業平台自行下載或更新所需的應用軟體程式，讓手機更具實用性。
隨著科技迅速發展，使用者可以隨時隨地利用手機的照相功能來拍攝照片，而隨著手機記憶卡容量愈來愈大，可存放的照片數量愈多，相對的使用者在整理照片上所需花費的時間也愈多，於是本論文將實作一個針對手機裡的照片自動作分類與剪裁處理的系統，讓使用者能更方便且快速的處整理手機照片，並增加搜尋照片的效率。
在本論文中，我們使用兩階段式的分類架構來實現系統的圖片分類功能，第一階段採用Android的人臉偵測方式偵測圖中的人臉區塊，並針對人臉區塊作膚色偵測的輔助判斷，先分類出手機圖片中包含有人物的圖片。第二階段我們提出利用圖片經過邊緣偵測後的二值化影像被影像侵蝕的比例去分類出建築物和風景圖片。最後以圖片分類時經過特徵分析的數據來訂定ROI(Region-of-Interest)的範圍，並針對ROI(Region-of-Interest)對圖片作剪裁處理。
本論文將實作一個適合嵌入式系統使用的較低運算複雜度(low-complexity)之影像剪裁/分類處理系統，並porting到Google Android的手機平台上，依open source的精神提供完整的程式給社群使用。

Cell phone is one of the most general consumer electronic products since modern times. Besides the call or SMS functions of the traditional cell-phone, the user can download and upgrade the applications through the Android Platform. The feature will make the cell-phone more practical.
With the rapid development of technology, user can take pictures with the cell-phone whenever and wherever. As memory card capacity of the cell-phone being larger and larger, more photos can be deposited. For this reason, user will waste much time to manage photos, and then we will implement a system that classifies and crops photos automatically in the cell-phone. User will manage photos faster and more convenient.
In the thesis, we implement the image classification function by two-stage classification process. In the first stage, we use face detect function in Android to detect face block region, and skin detection to judge possible miscarriage cases. We use this approach to find out photos with characters. In the second stage, we use masks to erode binary image after edge detection, and separate the photos with buildings and landscape photos by erosion rates. Finally, we use the data that analyzed through the image classification function to set the ROI(Region-of-Interest) range, and then crop the picture to ROI.
We will develop an appropriate digital image cropping/classification system with low complexity on the Google Android Platform. Then the complete program will be provided based on the spirit of open source.

[1] Gianluigi Ciocca, Claudio Cusano, Francesca Gasparini, Raimondo Schettini, “Self-Adaptive Image Cropping for Small Displays”
[2] Andrew Nusca (20 August 2009). "Smartphone vs. feature phone arms race heats up; which did you buy?". ZDNet.
[3] Windows Mobile 6.5, http://www.funddj.com/KMDJ/wiki/wikiviewer.aspx?keyid=ba9e989c-986d-4c3f-a82b-133503f8df01
[4] Android-Open Handset Alliance, http://en.wikipedia.org/wiki/Open_Handset_Alliance
[5] Android Developers, http://developer.android.com/intl/zh-TW/index.html
[6] What is Android? http://developer.android.com/intl/zh-TW/guide/basics/what-is-android.html
[7] H., Rowley, S., Baluja, T., Kanade, “Neural Network-Based Face Detection”, IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 20,(1), (1998) 23-28.
[8] M.H., Yang, D.J., Kriegman, N., Ahuja, “Detecting Faces in Images: A Survey”, IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 24(1), (2002) 34-58.
[9] P., Viola, M.J., Jones, “Robust real-time face detection”, International Journal of Computer Vision, vol. 57, pp. 137-154, 2004.
[10] Karin S., and Ioannis P. (1998), “A novel method for automatic face segmentation, facial feature extraction and tracking,” Signal Processing: Image Communication, Vol. 12, pp. 263-281
[11] R.C. Gonzalez and R.E. Woods, Digital Image processing(2nd), Prentice Hall, 2002.
[12] Feature Detectors – Sobel Edge Detector, http://homepages.inf.ed.ac.uk/rbf/HIPR2/sobel.htm
[13] L., Chen, X., Xie, X., Fan, W., Ma, H.J., Zhang, H.Q., Zhou, “A visual attention model for adapting images on small displays”, Multimedia Systems, vol. 9, pp. 353–364, 2003.
[14] L.Itti, C.Koch. Computational modeling of visual attention. Nature Reviews Neuroscience, 2001, 2(3): 194-203.
[15] http://www.sznews.com/photo/content/2009-02/01/content_3544410.htm