透過您的圖書館登入
IP:54.234.124.70
  • 學位論文

結合互資訊與蟻群覓食演算法從事影像套合之研究

Image Registration Using an Ant Colony Foraging Algorithm with Mutual Information

指導教授 : 張恆華

摘要


影像套合對於工程上的研究與醫療診斷用途上是相當重要的,其目的為對多張影像處理,並將個別的資訊顯示在套合後的影像。現存影像套合的方法相當多樣化,本論文敘述應用蟻群最佳化演算法來從事影像套合之研究。基本的蟻群最佳化演算法包含有路徑的選擇規則及費洛蒙更新規則,本研究嘗試使用蟻群最佳化演算法的基礎概念,並改良此蟻群最佳化演算法求出每次螞蟻移動方向並且計算螞蟻尋找食物的移動距離,進而求得每次套合的形變量,並結合雙線性內插法得到影像套合的結果。使用蟻群最佳化演算法可以省略黏性流體影像套合方法中,求解非線性偏微分方程式的繁複計算。本研究所使用的蟻群最佳化演算法依賴蟻群之覓食習性,再結合資訊理論中熵的概念,進而求得互資訊,以提高影像套合的精確度,並縮短影像套合所需的時間。我們使用了大量不同的影像包含空拍圖及醫學影像等來評估此一新方法。實驗結果證實本研究所提出的方法可有效解決多種不同影像套合的問題,而套合結果也相當的準確。在與黏性流體方法比較後顯示,本研究所提出的方法不僅有較高的相關係數,而且花費較少的執行時間。本論文所提之方法在多種不同種類的影像套合應用中具有相當的潛力。

並列摘要


Image registration is very important for a wide variety of image processing applications in engineering and medicine. It provides lots of image information for further analysis in many fields. There are many image registration methods being proposed. This thesis describes a new image registration algorithm using an ant colony optimization (ACO) approach. There are two fundamental properties in the proposed ACO process: the probabilistic transition and the pheromone update. We used the ACO algorithm to solve the direction and distance of advancement and combined linear interpolation to transform images. Thanks to the efficient ACO, complex calculation such as solving the Navier–Stokes partial differential equation is not necessary. The entropy condition in information theory was introduced to obtain interactive information in order to improve registration accuracy and reduce processing time. A wide variety of images including aerial images and medical images were used to evaluate this new method. Experimental results indicated that the proposed method efficiently performed registration and provided high accuracy. Comparing to the viscous fluid model method, our algorithm produced higher correlation coefficient scores but also spent less computation time. We believe that our algorithm is of potential in many image registration applications.

參考文獻


[27] 楊郁仙, “基於螞蟻演算法與路口延滯時間之最短時間路徑規劃,” 中興大學資訊科學與工程學系所學位論文, pp. 1-49, 2013.
[10] R. C. Gonzalez, and R. E. Woods, Digital image processing. 2002: publishing house of electronics industry, 2002.
[25] M. Dorigo, V. Maniezzo, A. Colorni, and V. Maniezzo, “Positive feedback as a search strategy,” 1991.
[1] J. Modersitzki, Numerical methods for image registration: Oxford university press, 2003.
[2] A. K. Jain, Fundamentals of digital image processing: prentice-Hall Englewood Cliffs, 1989.

延伸閱讀