基于人臉子區(qū)域加權(quán)和LDA的表情識(shí)別算法
發(fā)布時(shí)間：2018-12-07 10:56
【摘要】：人臉表情是人們?nèi)粘贤ㄖ凶钪匾囊环N表現(xiàn)特征,對(duì)于情感計(jì)算的研究發(fā)展具有重要意義。針對(duì)靜態(tài)表情圖像,單一的整體模版匹配方法特征維數(shù)較高,含有較多無(wú)關(guān)區(qū)域特征,因此很難獲得較好的識(shí)別效果。本文從幾何置信區(qū)域角度出發(fā),采用本文給出的基于幾何先驗(yàn)的加權(quán)策略進(jìn)行特征融合,結(jié)合線性判別分析算法進(jìn)行研究,獲得了較好的效果。本論文的主要工作如下:(1)給出一種基于子區(qū)域(置信區(qū)域)和多特征的加權(quán)融合特征提取算法。針對(duì)檢測(cè)區(qū)域存在較多與人臉部分不相關(guān)的區(qū)域,通過(guò)給出基于幾何先驗(yàn)的裁剪策略作用于檢測(cè)區(qū)域,得到更加精確的人臉及其子區(qū)域。針對(duì)人臉圖像存在表情無(wú)關(guān)區(qū)域并且單一特征描繪不準(zhǔn)確的特點(diǎn),采用Gabor小波對(duì)人臉區(qū)域進(jìn)行特征提取,HOG對(duì)置信區(qū)域進(jìn)行特征提取,通過(guò)研究置信區(qū)域在人臉表情中的先驗(yàn)信息(敏感度)并且實(shí)驗(yàn)加以論證,最終給不同的置信區(qū)域設(shè)置相應(yīng)權(quán)值,得到加權(quán)融合特征。在多個(gè)數(shù)據(jù)集上進(jìn)行實(shí)驗(yàn),驗(yàn)證了本文算法的有效性。(2)給出一種基于類內(nèi)散度矩陣修正的改進(jìn)LDA算法。針對(duì)初步降維特征缺少判別特性的缺點(diǎn),通過(guò)在類內(nèi)散度矩陣中引入余弦相似度信息,將每類樣本向量與其均值向量之間的夾角余弦值通過(guò)線性變換加權(quán)乘到對(duì)應(yīng)協(xié)方差矩陣中,獲得更好的類內(nèi)聚合度和類間離散度。在多個(gè)數(shù)據(jù)集上進(jìn)行實(shí)驗(yàn),驗(yàn)證了本文改進(jìn)算法的有效性。(3)構(gòu)造一種GRNN神經(jīng)網(wǎng)絡(luò)分類器首次應(yīng)用于人臉表情識(shí)別領(lǐng)域。針對(duì)傳統(tǒng)分類器對(duì)小樣本非線性數(shù)據(jù)擬合的局限性,通過(guò)對(duì)人臉表情數(shù)據(jù)特點(diǎn)的分析,構(gòu)造一種GRNN分類器嵌入表情識(shí)別算法中,將融合特征作為網(wǎng)絡(luò)的輸入,經(jīng)過(guò)模式層和求和層之后完成訓(xùn)練。在多個(gè)數(shù)據(jù)集上進(jìn)行實(shí)驗(yàn),驗(yàn)證了本文算法的有效性。
[Abstract]:Facial expression is one of the most important expression features in people's daily communication, which is of great significance to the research and development of emotional computing. For the static facial expression image, the single integral template matching method has higher feature dimension and more independent region features, so it is difficult to obtain a better recognition effect. In this paper, from the point of view of geometric confidence region, the feature fusion based on geometric priori weighting strategy is adopted, and the linear discriminant analysis (LDA) algorithm is used to study the feature fusion, and good results are obtained. The main work of this paper is as follows: (1) A weighted fusion feature extraction algorithm based on sub-region (confidence region) and multi-feature is proposed. Because there are many regions which are not related to the face part in the detection region, a geometric priori based clipping strategy is given to the detection region to obtain more accurate face and its sub-regions. In view of the feature of facial expression independent region and inaccurate description of single feature in face image, Gabor wavelet is used to extract the feature of face region, and HOG is used to extract the feature of confidence region. By studying the priori information (sensitivity) of the confidence region in the facial expression and proving it experimentally, the weighted fusion feature is obtained by setting the corresponding weights for the different confidence regions. Experiments on multiple datasets show the effectiveness of the proposed algorithm. (2) an improved LDA algorithm based on the correction of the intra-class divergence matrix is proposed. In view of the lack of discriminant characteristics in the preliminary dimensionality reduction feature, the cosine similarity information is introduced into the intra-class divergence matrix. The angle cosine value between each class of sample vector and its mean vector is weighted by linear transformation to the corresponding covariance matrix to obtain a better degree of intra-class aggregation and inter-class dispersion. Experiments on multiple datasets show the effectiveness of the improved algorithm. (3) A GRNN neural network classifier is first applied to facial expression recognition. Aiming at the limitation of the traditional classifier to fit the small sample nonlinear data, a GRNN classifier embedded in the facial expression recognition algorithm is constructed by analyzing the features of the facial expression data, and the fusion feature is taken as the input of the network. After the mode layer and summation layer completed the training. Experiments on multiple datasets show the effectiveness of the proposed algorithm.
【學(xué)位授予單位】：大連海事大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41
本文編號(hào)：2367089
[Abstract]:Facial expression is one of the most important expression features in people's daily communication, which is of great significance to the research and development of emotional computing. For the static facial expression image, the single integral template matching method has higher feature dimension and more independent region features, so it is difficult to obtain a better recognition effect. In this paper, from the point of view of geometric confidence region, the feature fusion based on geometric priori weighting strategy is adopted, and the linear discriminant analysis (LDA) algorithm is used to study the feature fusion, and good results are obtained. The main work of this paper is as follows: (1) A weighted fusion feature extraction algorithm based on sub-region (confidence region) and multi-feature is proposed. Because there are many regions which are not related to the face part in the detection region, a geometric priori based clipping strategy is given to the detection region to obtain more accurate face and its sub-regions. In view of the feature of facial expression independent region and inaccurate description of single feature in face image, Gabor wavelet is used to extract the feature of face region, and HOG is used to extract the feature of confidence region. By studying the priori information (sensitivity) of the confidence region in the facial expression and proving it experimentally, the weighted fusion feature is obtained by setting the corresponding weights for the different confidence regions. Experiments on multiple datasets show the effectiveness of the proposed algorithm. (2) an improved LDA algorithm based on the correction of the intra-class divergence matrix is proposed. In view of the lack of discriminant characteristics in the preliminary dimensionality reduction feature, the cosine similarity information is introduced into the intra-class divergence matrix. The angle cosine value between each class of sample vector and its mean vector is weighted by linear transformation to the corresponding covariance matrix to obtain a better degree of intra-class aggregation and inter-class dispersion. Experiments on multiple datasets show the effectiveness of the improved algorithm. (3) A GRNN neural network classifier is first applied to facial expression recognition. Aiming at the limitation of the traditional classifier to fit the small sample nonlinear data, a GRNN classifier embedded in the facial expression recognition algorithm is constructed by analyzing the features of the facial expression data, and the fusion feature is taken as the input of the network. After the mode layer and summation layer completed the training. Experiments on multiple datasets show the effectiveness of the proposed algorithm.
【學(xué)位授予單位】：大連海事大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41
【參考文獻(xiàn)】
相關(guān)期刊論文 前3條
1 薛雨麗;毛峽;張帆;;BHU人臉表情數(shù)據(jù)庫(kù)的設(shè)計(jì)與實(shí)現(xiàn)[J];北京航空航天大學(xué)學(xué)報(bào);2007年02期
2 ;Person-independent expression recognition based on person-similarity weighted expression feature[J];Journal of Systems Engineering and Electronics;2010年01期
3 王大偉;周軍;梅紅巖;張素娥;;人臉表情識(shí)別綜述[J];計(jì)算機(jī)工程與應(yīng)用;2014年20期
相關(guān)博士學(xué)位論文 前1條
1 楊章靜;基于鄰域結(jié)構(gòu)的特征提取及其在人臉識(shí)別中的應(yīng)用研究[D];南京理工大學(xué);2014年
，本文編號(hào)：2367089
本文鏈接：http://www.wukwdryxk.cn/shoufeilunwen/xixikjs/2367089.html
最近更新
教材專著
