數位影像為現今普遍使用的影像資料,對於如此快速的應用與大量使用的結果下,如何防止變造、破壞及破壞後的修復,也成為目前相關於影像的重要技術之ㄧ。如何讓儲存的影像或者是辨識影像的正確性也是ㄧ個重要課題。 本篇研究,著重於提供一個簡單快速的影像復原方法,利用模糊理論,做為本篇在經過浮水印影像復原後,針對無法復原的影像區塊,給予一個接近原色的演算方法。 在文中,選擇以浮水印的影像復原,做為第一次影像的復原,而以研究方法作為第二次影像復原,利用視覺和計算出來的PSNR值,來分析完成方法的效果及問題。在研究過程中,有點顛覆個人想法,讓自己的數學觀,由一個追求一個不是對就是錯的觀念,發展成為一個區間、程度,這是個人在研究過程中最大的收穫。
Digital images are a type of image format which are now in widespread used. Because digital images can be used in a huge scale and rapidly, it becomes an important part of image processing techniques to prevent them from being tampered and destroyed, also to recover when being destroyed. Therefore, what we emphasizetoday is how to store digital images and identify them accurately. In this study, I developed a simple and fast method of image recovering with fuzzy theory. This algorithm method gives image areas some colors that close to the original ones which could not be recovered through watermarking recovery. Hence, I chosed watermarking recovery to be the first-time image recovering and used the algorithm method just mentioned before to be the second-time. Then, analyzing the effectiveness of the results by our vision and a PSNR value and finding where the problem is. Through the research process, it totally reversed my view of mathematics from seeking for a dilemmatic answer to a multiple answer. This is what I gained most during research period!