傅楸善臺灣大學:資訊工程學研究所張庭瑄Chang, Ting-HsuanTing-HsuanChang2010-05-182018-07-052010-05-182018-07-052008U0001-1606200815074300http://ntur.lib.ntu.edu.tw//handle/246246/183642紅眼消除的技術是為了減少照片中的不協調的地方。此種技術能讓使用者得到更好的照片列印成果。近年來數位相機與行動電話的普及，使得人們可以輕鬆的拍取大量的相片，其中不乏在暗處使用閃光燈所拍攝到含有紅眼現象的圖片，而如何消除這個影響，已經成為一個重要的課題。 在這篇論文裡面，我們提出使用快速臉部偵測與自動偵測紅眼並消除的方法，自動的將影像中紅眼的部位消除，同時增快處理速度與降低假警報出現的機率，呈現出一張更為完美的輸出照片。Red-Eye Reduction Technology is used for some artifacts in still images. This technology can help user get a better photo print. In recent years, digital cameras and camera phones are more popular. People can take many pictures easily. Some photos taken in dark with flashlight will have some red-eye artifact. This subject becomes more important nowadays. In this thesis, we propose a fast face detection and automatic red-eye reduction method, which eliminates red-eye effect automatically. Increasing the processing speed and decreasing the false alarm rate to produce a better photo print.口試委員會審定書 i 謝 ii要 iiibstract ivhapter 1 Introduction 1.1 Motivation 1.2 Cause of Red Eye 2hapter 2 Related Work 4.1 Hardware solutions 4.2 Software Solutions 6hapter 3 Face Detection Method 13.1 Face Detection in Color Images 13.2 AdaBoost and Haar-like Filter 19hapter 4 Our Proposed Method 22.1 Accelerated Face Detection 22.2 Skin-Tone Verifying 26.3 Red-Eye Locating 27.4 Red-Eye Removal 30hapter 5 Experiment Results 33.1 Experiment environment 33.2 Experiment Result 34.3 Comparison 44hapter 6 Conclusion and Future Work 47.1 Conclusion 47.2 Future Work 47eference 49application/pdf5163581 bytesapplication/pdfen-US人臉偵測紅眼消除膚色偵測face detectionred-eye reductionskin-tone detectionthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/183642/1/ntu-97-R95922102-1.pdf