查詢結果分析
來源資料
頁籤選單縮合
題 名 | 2-D Binary Image Recognition using Grey Relation Analysis=二維黑白影像辨識--使用灰關聯分析 |
---|---|
作 者 | 周昌民; 蔣東建; 王俊勝; | 書刊名 | 健行學報 |
卷 期 | 19:1 1999.12[民88.12] |
頁 次 | 頁1-10 |
分類號 | 312.84 |
關鍵詞 | 灰關聯分析; 物體辨識; Grey relation analysis; Pattern recognition; |
語 文 | 英文(English) |
中文摘要 | 本文提出了一個灰關聯分析在物體辨識上的應用。文中所辨識的物體是以二維黑白影像為主,總共包含了兩個部分:特性擷取以及物體辨識。在特性擷取的部分,我們採用了影像在X軸及Y軸上投影曲線的傅立葉級數之係數做為代表一個物體的特徵值。而在物體辨識的部分,則是以灰關聯分析模型做為辨識物體的決策系統。 |
英文摘要 | This paper presents a possible application of Grey Relation Analysis Model for pattern recognition. Patterns used for classification in our works are 2 dimensional binary (i.e. black-white) images. Our study consists of two parts: feature extraction and pattern classification. In the feature extraction phase, we choose the Fourier series coefficients of the image's X-axis and Y-axis projection curves as its features. In the pattern classification phase, a Grey Relation Analysis Model is built as the decision function for pattern recognition. We choose a 10-class Arabic numerals (0,1,2,...,9) data set for experiments. Each numeral is stored as a 256 × 256 image. For each image, we extract 8 Fourier series coefficients from its X-axis projection curve and 8 coefficients from its Y-axis projection curve. Thus, each image has 16 features for representation. When a noised unknown image is analyzed, we extract 16 features from this pattern by the same manner. Then, the Grey Relation Analysis Model plays the role of decision function to figure out the Grey relational degrees between the unknown image and each of the original images for classification. This method achieves a 95% correct classification result when the SNR is 3. |
本系統中英文摘要資訊取自各篇刊載內容。