影像檢索技術在商標的搜尋應用上效果顯著，卻少有論文探討與商標類似之「印章」的檢索。本論文運用影像分割（segmentation）、特徵拮取（feature extraction）、檢索（retrieval）等方法之特性，以衝量守恆（Moment preserving）、哲尼克(Zernike moments, ZMs)、迦瑪函數(Gama function) 實現印章檢索功能。有關影像前處理之方法運用，為了取得較多的衝量訊息，以供特徵擷取之用，因此捨棄平滑法、中值法而採用“衝量守恆法”。由於ZMs易受影像中雜訊的影響，我們先找出影像形狀中心（shape gravity）作為原點，並找出最小外接圓，藉此消除雜訊影像並將形狀作初步的正規化。最後再將特徵影像映射至機率空間以加快檢索速度。上述檢索機制並非唯一，為了檢驗本論文所提方法之效果，將樣本影像分為十類，進行精確度及回收率（precision and recall）之測試，所得的檢索精確度至少在90%以上，因此，以此方法取代人工之「字義檢索」（text base image retrieval）確實可行。

關鍵字：影像檢索（retrieval）、衝量守恆、形狀中心、徑向角向基底、最小外接圓、精確度及回取率

Recognition by image contents has been studied for years. Practical processing steps to make it feasible comprise preprocessing, feature extraction and matching. Many applications of image recognition or retrieval can be found, however, few of them are concerned about the seal recognition. Actually, seal is an essential identification emblem in Chinese culture. We proposed to fulfill auto-accessing of seal cluster based on shape image retrieval method. In the preprocessing step, the moment preserving method is used to segment the seal from scanned images. To represent the seal feature with numerical values, the size and rotation invariant descriptors, Zernike moments, are extracted from the segmented seal image. A shape normalization process was proposed to make these descriptors robust to noises. The final query is just a process to measure distances between features of the query image and images in the database. With one scanned seal image, the user can easily identify whether the seal is fake or not. A seal includes some Chinese characters and it’s not just a trademark. To retrieve them automatically, quickly and correctly is important in oriental societies. Simulations demonstrate excellent retrieval performance based on the proposed methods.

[1] 王中山, “Using wavelet for license plate detection,” 中山大學機械與電機工程學系碩士論文, June 2004.
[2] Wen-Hsiang Tsai, “Moment preserving thresholding: a new approach,” Computer Vision, Graphics & Image Processing, vol. 29, no. 3, pp. 377-393, 1985.
[3] Michael Reed Teaque, “Image analysis via the general theory of moment,” J. Opt. Soc. Am., vol. 70, no. 8, pp. 920-930, Aug. 1980.
[4] Alireza Khotanzad and Yaw Hua Hong, “Rotation invariant image recognition using features selected via a systematic method,” Pattern Recognition, vol. 23, no. 10, pp. 1089-1101, 1990.
[5] S.-C. Pei and C.-N. Lin, “Image normalization for pattern recognition,” Image and Vision Computing, vol. 13, no. 10, pp. 711-723, Dec. 1995.
[6] S-.C. Pei and T.-C. S., “Normalization and recognition of affine-deformed image via moment-based technique,” National Taiwan University Electric Engineering Master graduation thesis, 1997.
[7] 陳子昂, 智慧型中文輸入---漢字識別、語音識別, 工業技術研究院電通所出版, 1994.
[8] 張真誠、黃國峰、陳同孝,電子影像技術, 松崗圖書, 2000.
[9] 張真誠、黃國峰、陳同孝,數位影像處理技術, 旗標圖書, 2003.
[10] 鍾國亮,資料壓縮的原理與應用,全華圖書, 2003.
[11] Sergios Theodoridis and Konstantinos Koutroumbas, “Pattern recognition,” Elsevier Acadmicc Press, 2nd ed., 2003.
[12] Peyton Z Peebles JR, “Probability, random variables and random signal principles,” McGRAW-HILL, 4th ed., 2001.
[13] Miroslaw Bober, “MPEG-7 visual shape descriptors,” IEEE Trans. Circuits & Systems for Video Technology, vol. 11, no. 6, June 2001.
[14] Daniel Keysers et al, “Adaptation in statistical pattern recognition using tangent vectors,” IEEE Transaction on Pattern Analysis and Machine Intelligence, vol. 26, no. 2, Feb. 2004.
[15] Kinh Tieu and Paul Viola, “Boosting image retrieval,” IEEE Conference on Computer and Pattern Recognition, 2000.
[16] Jung-Hsien Chiang, and Pei-Yi Hao, “A new kernel-based fuzzy clustering approach: support vector clustering with cell growing,” IEEE Trans. Fuzzy System, vol. 11, no. 4, August 2003.
[17] Yixin Chen and James Z Wang, “Support vector learning for fuzzy rule-based classification systems,” IEEE Trans. Fuzzy Systems, December 2003.
[18] 林宸生, 數位信號--影像與語音處理,全華圖書,修訂二版, 2003.
[19] J.-J. Chen, C.-Y. Liu, “A universal query mechanism for similarity retrieval based on information in image databases,” Int. Conf. Acoustic Speeches & Signal Processing, vol.3, pp. 3676-3679, 2002.
[20] Thomas Sikora, “The MEPEG-7 visual standard for content description—an overview,” IEEE Trans. Circuits & Systems for Video Technology, vol. 11, no. 6, 2001.
[21] P. Matsakis et al, “The Use of force histograms for affine-invariant relative position description,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 1, 2004.
[22] N. Otsu, “A threshold selection method from gray level histogram,” IEEE Trans. Systems, Man, and Cybernetics, SMC-8, pp. 62-66, 1978.
[23] J. N. Kapur, P. K. Sahoo, and A. K.C. Wong, “A new method for gray-level picture thresholding using the entropy of the histogram,” Computer Vision, Graphics & Image Processing, 29(3), pp. 273-285, 1985.
[24] R. Gupta and P. Evripidou, “Design and implementation of an efficient general purpose median filter network,” Digital Signal Processing, vol. 13, pp. 64-72, 1993.