多視角視網膜影像對位之研究

近年來影像處理技術應用於各種領域上，例如在醫療的領域上常扮演著輔助的角色。影像對位的技術對於醫療影像的分析也是目前被重視的課題。本文注重於尋找完整的視網膜血管特徵以提升後續對位上的特徵定位準確度，而影像強化(image enhancement)在醫學影像上一直為一個重要的應用。本研究需要克服因視網膜影像會含有細微血管特徵不明顯，影像色塊差異，旋轉變形問題以及因視角不同導致血管重疊區域大小等問題。因此本研究利用Gabor濾波器以及區域篩檢血管特徵，最後結果可以看到本研究細微血管特徵的提取上，相較於Canny提取更為精確。在尋找特徵點方面，利用加速穩檢特徵(Speeded Up Robust Features, SURF)演算法來尋找相鄰兩張不同視角視網膜的特徵點，並配合 RANSAC (Random Sample Consensus)演算法，篩去非正確的匹配，接著以迭代方式求出其轉換矩陣，最後將相鄰的兩張視網膜影像進行疊合，以求得更精準的影像資訊，實驗結果顯示，本研究所提出的特徵提取上可以增進定位精度且對位演算法具有一定的精準度。關鍵字：影像對位、影像強化、Canny、SURF、 RANSAC

關鍵字

影像強化；影像對位

並列摘要

Image processing technology increasingly widely used in various fields and plays a supporting role. For example, Image registration is also an important topic in medical image analysis. In this paper, there have some problems, like the image color is different, lost some subtle features or the image rotation problems. So we focus on how to detect the complete blood vessels and achieve more accurate matching. First, we use Gabor filter to detect edge and the adaptive method to detect edge. Second, we focus on image registration method which consists of a feature-based registration approach using speeded up robust features (SURF) and random sample consensus algorithm (RANSAC), to automatically registering two adjacent retinal images from a series of microscopic image sequence to achieve more accurate matching. The experimental results show that the proposed method in this study has promising results. Keywords: Image registration、Image enhancement、SURF、RANSAC

並列關鍵字

RANSAC ； Canny

參考文獻

A.T. Azar, S.A. El-Said. (2014), Performance Analysis of Support Vector Machines Classifiers in Breast Cancer Mammography

Bay, H., Tuytelaars, T., & Van Gool, L. (2006). Surf: Speeded up robust features.

risk index: Automated glaucoma detection from color fundus images. Medical

Image Analysis, 14(3), 471-481.

Bovik A C，Clark M，Geisler W S.(1990). Multichannel texture analysis using localized spatial filters[J].IEEE Trans Pattern

國際替代計量

多視角視網膜影像對位之研究

未授權

主題瀏覽