基於快速傅立葉轉換之立體圖像資料庫探討

A Study of 3D Trademark Image Database Based on FFT

賴亮秀(LIANG-HSIU LAI)

指導教授 ： 李朱慧

朝陽科技大學/資訊學院/資訊管理系/碩士(2017年)

摘要

立體商標圖像隨著經濟貿易國際化以及網路時代服務的來臨，逐漸受到許多用戶用來做為形象商標，使用者會在網路上或是實體店面來使用立體商標圖像。因此在合法使用前，立體商標圖像是需要申請註冊的，目前很多國家都可以進行立體商標圖像的申請註冊，在申請註冊之前，必須要檢視與目前已完成註冊立體商標圖像的相似性是現在很重要的工作，因此，如何有效率且迅速的檢索立體商標圖像是很重要的議題。本研究使用臺灣經濟部智慧財產局(IPO)所提供公開的立體商標資料庫平台提出一個有效率的檢索方式，目前在臺灣一組立體商標圖像是使用許多張 2D 的圖像來代表。本研究對於檢索立體商標圖像，提出了用快速傅立葉轉換(Fast Fourier transform ,FFT)的技術來提取立體商標圖像的特徵，實驗中將多張的 2D 圖像的特徵結合與提取，使檢索的結果有效且迅速。

關鍵字

圖像檢索 ； 快速傅立葉轉換 ； 相似度測量 ； 3D立體商標

並列摘要

With the diversity of the trademark, three-dimensional (3D) trademark has gradually been used by many companies. For the legitimate use, 3D trademark is required to be registered. In Taiwan, the registration application is proved by Intellectual Property Office under the Ministry of Economic Affairs. The 3D trademark is useful to identity characteristic products or company on the Internet or the physical store. At present, 3D trademark can be registered in many nations. However, it is must be made sure unique in the 3D trademark gallery before registration. Similarity measurement of 3D trademark is an important task and how to efficiently searching is an interesting issue. This is an efficient searching method for 3D trademark gallery in Taiwan’s Intellectual Property Office is provided. A 3D trademark is represented by multiple 2D images in the gallery. Fast Fourier transform (FFT) is used to extract the feature, and the effective search mechanism is established by using the properties of multiple 2D images. The experiment showed the method was efficient.

並列關鍵字

Image retrieval ； Fast Fourier transform (FFT) ； similarity measurement ； 3D trademark

參考文獻

[1] 李朱慧*、林金樹(2013)，”3D影像查詢系統，”資訊科技國際研討會，朝陽科技大學，April 26-27.(特徵權重)

[4] Kallianpur, Akshay K., M. V. Bharath, and K. Manikantan(2015), “Digital image watermarking using optimized transform-domain approach,” IEEE UP Section Conference on Electrical Computer and Electronics (UPCON), pp.1-6.

[6] Cheng-Chieh Chiang, et al(2009), “Region-based image retrieval using color-size features of watershed regions.” Journal of Visual Communication and Image Representation, Vol. 20, No. 3, pp. 167-177.

[7] Park Chun-Su(2015), “2D discrete Fourier transform on sliding windows.” IEEE Transactions on Image Processing, Vol.24, No.3, pp.901-907.

[8] Barna Csuka, et al(2016), “Comparison of Signal Processing Methods for Calculating Point-by-point Discrete Fourier Transforms,” IEEE 26th International Conference Radioelektronika (RADIOELEKTRONIKA), pp.217-221.