【正文】 [19]王一丁,李鎏,楊虹. 使用漢明距離約束的LBP人臉識別算法[J]. 計算機工程與應(yīng)用,2009,27:188190+236.[20]趙葛宏. 基于小波多層次LBP算法的生物特征識別與分類[D].昆明理工大學(xué),2012.[21]毛科. 基于局部二值模式的人臉識別算法研究[D].華南理工大學(xué),2010.[22]王凱. 基于圖像紋理特征提取算法的研究及應(yīng)用[D].西南交通大學(xué),2013.[23]趙懷勛,徐鋒,陳家勇. 基于多尺度LBP的人臉識別[J]. 計算機應(yīng)用與軟件,2012,01:257259+279.[24]王憲,張彥,慕鑫,張方生. 基于改進的LBP人臉識別算法[J]. 光電工程,2012,07:109114.[25]黃非非. 基于LBP的人臉識別研究[D].重慶大學(xué),2009.[26]周凱. 基于局部二值模式的人臉識別方法研究[D].中南大學(xué),2009.[27]閆偉紅. 基于LBP統(tǒng)計特征的人臉識別方法研究[D].安徽大學(xué),2010.附 錄 英文原文：Feature local binary patterns with application to eye detectionAbstractThis paper presents a new Feature Local Binary Patterns (FLBP) method that encodes the information of both local texture and features. The features are broadly defined by, for example, the edges, the Gabor wavelet features, the color features, etc. Specifically, a binary image is first derived by extracting feature pixels from a given image, and then a distance vector field is obtained by puting the distance vector between each pixel and its nearest feature pixel defined in the binary image. Based on the distance vector field and the FLBP parameters, the FLBP representation of the given image can be formed. In contrast to the original The LBP with Relative Bias Thresholding (LRBT)1. IntroductionThe Local Binary Patterns (LBP) method, which defines a grayscale invariant texture description by paring a center pixel with its neighbors, is a popular method for texture analysisand Gong et al.[7], the color features (iii) the FLBP method displays superior representational power and flexibility to the LBP method due to the introduction of feature pixels as well as its parameters。[13]introduced Local Ternary Patterns (LTP), for face recognition. Liu and Liu[21].The mon features in the distinctive featurebased approaches include edge, intensity of iris, as well as color distribution. Feng and Yuenand Kawato and Tetsutaniproposed a method for locating the iris of an eye using both intensity and edge information. Other methods[34],這個特征是由廣義的，例如，邊緣，Gabor小波特征，色彩特征，等。新LRBT方法有助于與其他特征像素提取方法相比提高了FLBP眼檢測性能。特別是，首先，F(xiàn)LBP方法顯著于LBP方法提高了在這兩個眼檢測率和眼中心定位精度方面。proposed a method for detecting eye and mouth using distance vector field. Recently, Chen and Liuproposed an approach that uses a morphological method for extracting an eye strip, where the iris is located through template matching by means of an adaptive half circle template.The photometric appearancebased approaches usually collect a large amount of training data representing the eyes of different subjects, with different face orientations and under different illumination conditions. A classifier or regression model is then constructed for eye detection. The Eigen analysis has been applied in eye detections sclera, and a Gaussian filter for detecting the dark circle of the iris. The nonlinear filter is used to detect the left and right corners of an eye in a color image. Kawaguchi and Rizonextended the VPF to a generalized projection function (GPF). Their experiments show that the hybrid projection function, which is a special case of GPF, is better than VPF, while VPF is better than the integral projection function. Kawato and Ohya[20]introduced Local Quantized Patterns (LQP), a generalization that uses lookuptable based vector quantization to code larger or deeper patterns. Tan and Triggs[12]Section ). The contributions of the paper are as follows:?A new FLBP method is presented. The FLBP encodes both local and feature information. In contrast to the original LBP that only pares a pixel with the pixels in its own neighborhood, the FLBP can pare a pixel with the pixels in its own neighborhood as well as in other neighborhoods. The FLBP generalize the LBP which can be considered as a special case of the FLBP. The FLBP is expected to perform better than the LBP approach for texture description and pattern recognition.?As the FLBP method encodes both local and feature information, the performance of FLBP depends on the extraction of the feature pixels. To improve FLBP performance, we present a new feature pixel extraction method, the LBP with Relative Bias Thresholding (LRBT) method.?For the application of FLBP on eye detection, experimental results using the BioID and FERET databases show that: (i) the FLBP method significantly improves upon the LBP method in terms of both eye detection rate and eye center localization accuracy。[6]introduced the concept of texture unit and texture spectrum. A texture unit of a pixel is represented by eight elements, which correspond to the eight neighbors in a 33 neighborhood with three possible values: 0, 1, 2. The three values represent three possible relationships between the center pixel and its neighbors: “l(fā)ess than”, “equal to”, or “greater than”. As a result, there are 38=6561 possible texture units in total. A texture spectrum of a region is defined by the histogram of the texture units over the region. The large number of possible texture units, however, poses a putational challenge. To reduce the putational burden, Ojala et al.老師對于學(xué)生總是默默的付出,盡管很多時候我們自己并沒有特別重視論文的寫作,沒有按時完成老師的任務(wù),但是老師還是能夠主動的和我們聯(lián)系,告訴我們應(yīng)該怎么樣修改論文,怎么樣按要求完成論文相關(guān)的工作。老師的意見總是很寶貴的，可以很好的指出我的資料收集的不足以及需要什么樣的資料來完善文章。歷經(jīng)了這么久的努力，終于完成了畢業(yè)論文。 作為實驗1中的set1部分訓(xùn)練樣本表 1 在Yale人臉庫中對LMCP和LBP進行實驗1的對比 （%）Set2Set3Set4Set5Ave圖 表 2 在Yale人臉庫中對LMCP和LBP進行實驗2的對比 （%）Set1Set2Set4Set5Ave圖 表 3 在Yale人臉庫中對LMCP和LBP進行實驗3的對比 （%）Set1Set2Set3Set4Set5Ave 基于ORL人臉庫的實驗我們抽取了0RL人臉庫中40個人,每人10張閣片共400張圖片，采用了 2張訓(xùn)練圖片8張測試圖片；4張訓(xùn)練圖片，6張測試圖片這兩組來進行了實驗,測試次數(shù)為400次。本文代碼是基于matlab中的M語言所寫，使用的編譯環(huán)境為Matlab 2010b。該數(shù)據(jù)庫中包含了分別屬于1040志愿者的99450張不同的人臉圖片。截至1997年,該人臉庫中總共包含了分別屬于1199人的14126張人臉圖片,人臉圖片分別是在不同的外部條件(光照,姿態(tài),表情,時間間隔)下進行拍照獲得。本文中的測試主要使用Yale人臉庫以及ORL人臉庫上進行。如果將不屬于這些取值范圍的 全部歸為另一種取值，那么可以將特征維度大約降低為原來的1/6 。圖中左右兩幅圖片表示圖像局部區(qū)域像素值，它們得到的LBP 值完全相同，但是卻忽略掉了十分重要的對比度信息，而對比度的不同恰恰是這兩個局部區(qū)域紋理