以自然影像為基礎之蝴蝶辨識系統

A Natural Image-based Butterfly Recognition System

吳明政(Ming-Jheng Wu)

指導教授 ： 劉震昌

國立暨南國際大學/科技學院/資訊工程學系/碩士(2011年)

摘要

網際網路的發達，改變了人們的生活形態，一般大眾在遇到問題的時候，他們的第一選擇便是在網際網路上求助。以蝴蝶辨識為例，蝴蝶類別的查詢己經不再侷限於實體圖鑑或是與專家的詢問。網路搜尋引擎提供了關鍵字查詢，但是蝴蝶影像難以用單純的文字表示，使用者無法下出適當的關鍵字得到良好的結果。
本論文研究以影像內容為基礎的蝴蝶辨識系統，系統能擷取含有自然背景的蝴蝶影像特徵，包括色彩直方圖，SIFT 局部特徵與局部仿射不變性部分三種特徵，藉由特徵比對的方式搜尋資料庫中相似的蝴蝶影像，達到蝴蝶類別辨識。亦研究使用影像物件性來框選影像中最有可能的物件區域。比較上述不同的方法，我們最高可以達到平均95.2% 的自然蝴蝶影像辨識率，即便這些自然蝴蝶影像是從網路收集，未經過人工的標記或是切割，並且影像背景充滿雜訊。

關鍵字

影像搜尋 ； 蝴蝶辨識 ； 局部仿射不變性部分 ； 影像物件性

並列摘要

Popularity of the Internet has dramatically changed our life. People are used to ask question on Internet when they have problems. For example of butterfly recognition, we are not limited to querying butterfly field guides or asking the experts. Search engines provide keyword search. However, it is hard to represent butterfly images using proper keywords.
The thesis studies content-based butterfly image recognition. Features are extracted from the natural images containing a butterfly. Three feature extraction methods are compared, including color histogram, SIFT, and semi-local affine part. Similar images in the database are ranked by feature matching. Category of the butterfly can be decided from the search result. The objectness measure is also studied to select the most likely window containing the object in the natural image. In the experiments, the dataset is composed of butterfly images from the Internet, which are not manually segmented and include cluttered background. The best recognition rate is 95.19%.

並列關鍵字

image search ； butterfly recognition ； semi-local affine part ； objectness

參考文獻

[1] M. Flickner, H. Sawhney, W. Niblack, J. Ashley, Q. Huang, B. Dom, M. Gorkani, J. Hafner, D. Lee, D. Petkovic, D. Steele, and P. Yanker, “Query by Image and Video Content : The QBIC System,” IEEE Computer, vol. 28, pp. 23-32, Sept. 1995.

[2] W. Niblack, R. Barber, W. Equitz, M. Flickner, E. Glasman, D. Petkovic, P. Yanker, C. Faloutsos, and G. Taubin, “The QBIC Project: Querying Images by Content using Color, Texture and Shape,” in Proceedings of the SPIE on Storage and Retrieval for Image and Video Databases, vol. 1908, pp. 173-187, 1993.

[3] S. Newsam, B. Sumengen, and B. S. Manjunath, “Category-Based Image Retrieval,” in Proceedings of the IEEE International Conference on Image Processing, vol. 3, pp. 837-842, Aug. 1996.

[4] M. N. Do and M. Vetterli, “Wavelet-Based Texuture Retrieval using Generalized Gaussian Density and Kullback-Leibler Distance,” IEEE Transactions on Image Processing, vol. 11, Feb. 2002.

[5] B. G. Prasad, K. K. Biswas, and S. k. Gupta, “Region-Based Image Retrieval using Integrated Color, Shape and Location Index,” Computer Vision and Image Understanding, vol. 94, pp. 193-233, 2004.