- 學位論文
植基於鏈結串列結構之影像邊緣偵測及其應用於資訊隱藏之研究
A Study on Edge Detection by Using the Chained Linked List Structure and Its Application for Information Hiding
摘要
邊緣偵測在數位影像處理與電腦視覺中是一項相當重要的基本處理技術，這是因為影像邊緣能夠有效地應用於影像分割、物件偵測等等領域。於2005年時，一個以最小擴張樹(MST)結構為基礎邊緣偵測演算法被發表，相較於梯度運算子(Gradient Operator)的邊緣偵測法，該演算法之可獲得較佳之處理結果。以像素值差分法(PVD)為基礎的資訊隱藏法發表於2003年，利用兩像素值之差分，可將機密資料藏於其中；其後，於2005年時，將像素值差分法與最不重要位元法(PVD&LSB)結合之新資訊隱藏法被提出，該方法改良自原本的像素值差分法，可使得資訊隱藏容量大幅提升。 在本研究中，我們提出了以串聯式鏈結串列結構為基礎的邊緣偵測演算法(CLL)，與最小擴張樹測邊法比較後，我們的演算法在時間效率方面具有較佳的執行速度，不但具有區分邊緣與非邊緣之能力，並能夠減弱影像中非邊緣區域之雜訊干擾。除此之外，利用此我們所提出之測邊法，將其進一步應用在資訊隱藏技術中，能夠動態地決定最不重要位元的位元數，使其具有高資訊隱藏量及高偽裝影像品質之能力。從實驗結果可得知，與其他類似的方法相互比較後，我們提出的方法能夠在相似的影像品質下，藏入更多的機密資料。
並列摘要
Edge detection is a very important technique in the domain of image processing and computer vision. Because edge detection is useful in applying to image segmentation and object detection. An edge detection method based on Minimum Spanning Tree (MST) was published in 2005. The edge detection of MST method can produce better image quality than gradient operators. An information hiding algorithm based on Pixel-value Differencing (PVD) was announced in 2005. It can hide the secret data within the difference of two neighboring pixels. Afterward a new information hiding algorithm was issued in 2005. The new algorithm is based on combine PVD and Least Significant Bit (LSB). It can increase plenty of capacities more than the PVD algorithm. This study provides an edge detection method by chained linked-list. After compared to MST method, the proposed of the present study method has better time efficiency in processing speed. And the proposed method can detect the edge and the area of plain, it can reduce noises in the area of plain. In addition, CLL can be applied to information hiding, called Dynamic LSB by Edge Detection (DLED). DLED can determine the bits of LSB dynamic to improve hiding capacity and image quality. In current research, it is found that DLED can hide more secret data than PVD and others.