標題: 影像與倒影之對稱偵測
Detection of Imagery Symmetry due to Optical Reflection
作者: 梁元宇
Liang, Yuan-Yu
林正中
Lin, Cheng-Chung
資訊科學與工程研究所
關鍵字: 對稱偵測;資訊梯度統計法;霍夫轉換;Symmetry Detection;Gradient Orientation Histogram;Hough Transform
公開日期: 2012
摘要: 本論文就影像與對稱倒影之對稱偵測課題加以探究並實作出可行有效之系統。其處理過程分成三個階段:1)影像特徵點擷取2)梯度資訊統計分析3)對稱軸及對稱點配對搜尋。 論文研究重點在於特徵點擷取階段,整合亮度對比與兩種特定梯度結構分佈類型所取得之特徵點,特別適用於影像與倒影之對稱偵測之用。
Study on the issues of detecting imagery symmetry due to optical reflection is presented in this thesis, based on which a system capable of effectively finding such reflective imagery symmetry was implemented. The key lies in the extraction of feature points for feeding into a quite standard symmetry detecting process. The extraction exploits specific patterns in gradient structure in conjunction with contrast such that feature points in favor of reflective imagery symmetry detection are extracted.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT079955533
http://hdl.handle.net/11536/50449
Appears in Collections:Thesis


Files in This Item:

  1. 553301.pdf

If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.