一个有效的闪光灯摄影增强法AnEfficientMethodforFlash.PDF

一個有效的閃光燈攝影增強法 An Efficient Method for Flash Photography Enhancement 吳俊霖 (Jiunn-Lin Wu), 洪夢韓(Meng-Han Hung) 中興大學 資訊科學系 {jlwu, s9256015}@cs.nchu.edu.tw 摘要 camera settings often results in camera motions. Since the flash/no-flash images do have significant differences due to the change in illumination, we 如果在偏暗的環境下要拍照的話，通常我們會 propose a robust image registration method based on 使用閃光燈，藉以補光以得到更多的影像細節，但 FFT is proposed for registration of such image pairs in 不幸的是這樣卻會破壞相片中的色彩。要是不使用 this paper. An image pair with moving objects is used 閃燈的話，雖然可以得到色彩真實的影像，但是其 to demonstrate the robustness of the proposed method. 需要較長的曝光時間或須以高ISO值拍攝，這會導 In this paper, we also figure out the problem in 致影像的模糊或產生嚴重雜訊。所以學者提出藉由 the conventional flash photography enhancement 結合一張沒有使用閃光燈與另一張有使用閃光燈 methods, they usually require a large amount of memories during the computation. It is a serious 的兩張相片來計算產生一張同時具有真實色彩與 computational burden when the size of the input 細節豐富的影像。 images is large. In this paper, we propose a new 本論文中，我們首先將提出針對有時使用閃光 method based on YSbSr color space to alleviate this 燈的影像序列的強健影像校正(image registration) problem. Experimental results show that the 法，其是基於快速傅立葉轉換，所以能有效克服影 advantages of the proposed method include: (1) it 像序列中光照差異大，甚至有移動物體的問題。接 saves up to 50% of the memo