基因類神經網路用於快速自動對焦法之研究

本論文主要目的為研究一個嶄新的被動式快速自動對焦方法，此方法主要分成兩個研究階段，第一階段首先是針對目前常用的離散餘弦轉換進行簡化，傳統清晰尺度的評判是將離散餘弦轉換後的係數全部加總，而對焦點則是清晰尺度中係數總和最大的影像，然而實際上離散餘弦轉換後的低頻係數主要受到影像內容所影響，而高頻係數一部分是由影像設備之雜訊所引起，因此本論文第一階段研究乃是利用基因演算法所提供之多方向搜尋的優點，分離出一個精簡的離散餘弦轉換係數區域，此區域除了能夠捨去低頻和高頻的影響外，亦可減少離散餘弦轉換的計算工作量，並且仍然有足夠的資訊能夠顯示出正確對焦點位置。本論文第二階段是在第一階段的精簡係數區域進行研究，主要目的是為了訓練出一個能夠藉由離散餘弦轉換係數預測對焦位置的類神經模型，本階段首先利用基因演算法作為搜尋媒介，挑選出兩組離散餘弦轉換係數範圍，將其當作類神經網路的輸入層，並以實際對焦點與類神經預測之對焦點誤差當作誤差訊號，之後利用誤差倒傳遞之訓練方式，使預測模型收斂，藉此訓練出能夠由前後影像之特定係數推導出對焦位置的類神經預測模型，達到快速對焦的能力。

關鍵字

離散餘弦轉換；基因演算法；類神經網路；自動對焦

並列摘要

In this study, the fast auto-focusing method using GA-based neural network prediction model by specific DCT coefficients is presented. There are two research stages in this method. In the first stage, the traditional measurement of AF by DCT has modifies by genetic algorithm to searching a range of DCT coefficient that can simply calculated without missing best-focusing position. In the second stage, it is to training a GA-based neural network prediction model. The main idea is using a neural network to approximate a prediction model that can predict the best-focusing position by two ranges of specific DCT coefficients of the first stage result in adjacent images. But there is no research about how selects the DCT coefficients. Besides, the how many neurons to use for the hidden layer is also a problem in neural network. Thus the GA search was combined with the neural network to solve these problems. At the finals, the experimental result shows that the proposed method can work efficiently.

並列關鍵字

auto-focusing (AF) ； neural network ； discrete cosine transform (DCT) ； Passive ； genetic algorithm (GA)

參考文獻

[1] C.M. Chen, C.M. Hong, H.C. Chuang, ”Efficient auto-focus algorithm utilizing discrete difference equation prediction model for digital still cameras,” IEEE Transactions on Consumer Electronics, vol. 52, no. 4, pp. 1135-1143, 2006.

[2] J.S. Lee, Y.Y. Jung, B.S. Kim, S.J. Ko, “An advanced video camera system with robust AF, AE, and AWB control,” IEEE Transactions on Consumer Electronics, vol. 47, no. 3, pp. 694- 699, 2001.

[3] J.H. Lee, K.S. Kim, B.D. Nam, J.C. Lee, Y.M. Kwon, H.G. Kim, “Implementation of a passive automatic focusing algorithm for digital still camera,” IEEE Transactions on Consumer Electronics, vol. 41, no. 3, pp. 449-454, 1995.

[4] Jie He; Rongzhen Zhou; Zhiliang Hong, ”Modified fast climbing search auto-focus algorithm with adaptive step size searching technique for digital camera,” IEEE Transactions on Consumer Electronics, vol. 49, no. 2, pp. 257-262, 2003.

[5] M. Subbarao, J.K. Tyan, “Selecting the Optimal Focus Measure for Autofocusing and Depth-From Focus,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 20, no. 8, pp. 864-870, 1998.

國際替代計量

基因類神經網路用於快速自動對焦法之研究

未授權

主題瀏覽