網點加密圖像光學擷取比對辨識研究

印刷圖像防偽技術除了在印刷材料與製程技術提昇改進外；亦可考慮印刷品進行數資料加密（Data Hiding）以提高其安全性。本研究針對紙張上印刷過網圖像進行加密與比對辨識實驗，將資料加密技術結合數位過網演算法製作網點加密圖像，並對圖像進行光學擷取後再進行資料辨識。旨在探討網點加密圖像在不同過網解析度（150/300LPI）下以不同光學資料擷取解析度（600/1200/2400DPI）對加密資料的比對識正確率高低，瞭解網點加密圖像比對辨識條件。
本研究印刷圖像辨識錯誤現象，在圖像過網解玎度150DPI時，以光學解析度1200或2400DPI擷取圖像其辨識錯誤次數最少可獲得最穩定辨識結果。將網點圖像予以加密複製於平面媒體上，有利於印刷品防止變造與安全保護。

In addition to improve the materials and the process for printing images, counterfeiting technology can be used for digital data hiding on prints to enhance their security. The objective of this research is focusing on the printing halfone image on papers to make an experiment on data hiding and recognition to combine the data hiding technology with digital halftoning algorithms to produce data hiding in halftone image and to process the data recognition after optical retrieving of images. In order to understand the most optimized condition for the recognition of data hiding in halftone image, we could discuss the influence of different optical data acquirement resolution (600/1200/2400DPI) under different halftoning resolution (150/300LPI) on the correct rate of the recognition of encrypted image data.
Using optical resolution 1200 or 2400 DPI to acquire the images with image halftoning resolution 150LPI, the error recognition phenomenon of this printing image research can obtain the steadiest recognized results with low-error-frequency. It is beneficial to encrypt halftone images on the plane media for the authentication and security protection of print matters.