【摘要】：人臉表情是人類交流的一種必不可少的方式,通過對(duì)人臉表情的研究可以探索人類的心理狀態(tài),進(jìn)而充分的了解人們的行為意圖。深度學(xué)習(xí)是一種特征學(xué)習(xí)方法,通過一些簡(jiǎn)單的非線性的模型把數(shù)據(jù)轉(zhuǎn)變成為更高層次的,更加抽象的表達(dá),可以解決語音處理、計(jì)算機(jī)視覺、自然語言處理等問題。本文利用深度學(xué)習(xí)來解決表情識(shí)別中的一些問題,并通過實(shí)驗(yàn)進(jìn)行了驗(yàn)證。本文的主要研究內(nèi)容如下:1.本文研究了多種類型的深度學(xué)習(xí)模型,按結(jié)構(gòu)可分為深度卷積神經(jīng)網(wǎng)絡(luò)、深度信念網(wǎng)絡(luò)、深度玻爾茲曼機(jī)、堆疊自動(dòng)編碼器和遞歸神經(jīng)網(wǎng)絡(luò)等。它們算法各異,且所適用領(lǐng)域也不相同。因此,選擇合適的深度學(xué)習(xí)模型是解決表情識(shí)別問題的關(guān)鍵。本文經(jīng)過對(duì)比論證發(fā)現(xiàn),深度卷積神經(jīng)網(wǎng)絡(luò)其獨(dú)特局部連接與權(quán)值共享機(jī)制可以解決表情識(shí)別中特征維數(shù)大,計(jì)算困難等問題。因此,本文選擇深度卷積神經(jīng)網(wǎng)絡(luò)作為本文的深度學(xué)習(xí)模型。2.針對(duì)靜態(tài)表情識(shí)別中提取特征會(huì)損失圖像原有的特征信息的問題,提出利用深度學(xué)習(xí)模型中的深度卷積神經(jīng)網(wǎng)絡(luò)實(shí)現(xiàn)表情特征提取。由于深度卷積神經(jīng)網(wǎng)絡(luò)避免了對(duì)圖像的復(fù)雜前期預(yù)處理,可以直接輸入原始圖像,它通過卷積與池化的共同作用進(jìn)行特征提取,不需要人為提取特征,并且網(wǎng)絡(luò)易于訓(xùn)練,比那種全連接的神經(jīng)網(wǎng)絡(luò)的泛化性能更好,因此將深度卷積神經(jīng)網(wǎng)絡(luò)應(yīng)用靜態(tài)表情識(shí)別中。3.針對(duì)動(dòng)態(tài)表情識(shí)別中系統(tǒng)抗干擾性差,運(yùn)算速度慢以及實(shí)時(shí)性差等問題,提出利用深度卷積神經(jīng)網(wǎng)絡(luò)實(shí)現(xiàn)動(dòng)態(tài)表情特征提取。由于動(dòng)態(tài)表情識(shí)別系統(tǒng)輸入的實(shí)時(shí)獲取的動(dòng)態(tài)人臉表情序列,與靜態(tài)表情識(shí)別不同,這需要系統(tǒng)實(shí)時(shí)將獲取到的人臉進(jìn)行存儲(chǔ)和識(shí)別。為了解決這一問題,利用Haar分類器進(jìn)行人臉檢測(cè),再引入深度卷積神經(jīng)網(wǎng)絡(luò),通過它構(gòu)建圖像的本質(zhì)特征,提取表情特征,最后運(yùn)用Softmax分類器實(shí)現(xiàn)表情分類。4.為了提高深度卷積神經(jīng)網(wǎng)絡(luò)的非線性表達(dá)能力,從而實(shí)現(xiàn)更好地表情特征提取,對(duì)網(wǎng)絡(luò)結(jié)構(gòu)進(jìn)行改進(jìn),利用深度連續(xù)卷積神經(jīng)網(wǎng)絡(luò)來實(shí)現(xiàn)表情識(shí)別。本文對(duì)深度卷積神經(jīng)網(wǎng)絡(luò)進(jìn)行改進(jìn),引用多層小尺度卷積代替單層大尺度卷積的思想,即利用2層連續(xù)卷積層來代替單層卷積層,提高網(wǎng)絡(luò)的非線性表達(dá)能力,再對(duì)網(wǎng)絡(luò)的激活函數(shù)及參數(shù)優(yōu)化方法做出調(diào)整,來提高網(wǎng)絡(luò)的表情特征擬合能力。
[Abstract]:Facial expression is an indispensable way of human communication. Through the study of facial expression, we can explore the psychological state of human beings, and then fully understand the behavior intention of human beings. Deep learning is a feature learning method. It can solve the problems of speech processing, computer vision, natural language processing and so on by transforming data into higher level and more abstract expression through some simple nonlinear models. In this paper, deep learning is used to solve some problems in expression recognition, and it is verified by experiments. The main contents of this paper are as follows: 1. In this paper, many kinds of deep learning models are studied, which can be divided into deep convolution neural network, depth belief network, depth Boltzmann machine, stacking automatic encoder and recurrent neural network. They have different algorithms and different fields of application. Therefore, choosing the appropriate deep learning model is the key to solve the problem of expression recognition. Through comparison and demonstration, it is found that the special local connection and weight sharing mechanism of deep convolution neural network can solve the problems of large feature dimension and difficult computation in facial expression recognition. Therefore, in this paper, the deep convolution neural network is chosen as the depth learning model of this paper. 2. In order to solve the problem that feature extraction in static expression recognition will lose the original feature information of image, this paper proposes to use depth convolution neural network in depth learning model to realize expression feature extraction. Because the deep convolution neural network avoids the complex pre-processing of the image, it can directly input the original image. It can extract the features through the joint action of convolution and pooling, and it does not need man-made feature extraction, and the network is easy to train. The generalization performance of the fully connected neural network is better than that of the fully connected neural network, so the deep convolution neural network is applied to static expression recognition. In order to solve the problems of poor anti-jamming, slow computing speed and poor real-time performance in dynamic expression recognition, a method of dynamic expression feature extraction based on deep convolution neural network is proposed in this paper. Because the real-time acquired dynamic facial expression sequence from the dynamic expression recognition system is different from the static facial expression recognition, it is necessary for the system to store and recognize the acquired face in real-time. In order to solve this problem, the Haar classifier is used for face detection, and then the deep convolution neural network is introduced to construct the essential features of the image, extract the expression features, and finally use the Softmax classifier to realize the expression classification. 4. In order to improve the nonlinear expression ability of the deep convolution neural network and achieve better expression feature extraction, the network structure is improved and the deep continuous convolution neural network is used to realize the expression recognition. In this paper, the deep convolution neural network is improved, and the idea that the multi-layer small-scale convolution replaces the single-layer large-scale convolution is introduced, that is, the two-layer continuous convolution layer is used to replace the single-layer convolution layer, and the nonlinear expression ability of the network is improved. Then the activation function and parameter optimization method of the network are adjusted to improve the expression feature fitting ability of the network.
【學(xué)位授予單位】：長春工業(yè)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10條

1 張友梅;張偉;;基于數(shù)據(jù)融合的表情識(shí)別[J];四川大學(xué)學(xué)報(bào)(工程科學(xué)版);2016年06期

2 何小飛;鄒崢嶸;陶超;張佳興;;聯(lián)合顯著性和多層卷積神經(jīng)網(wǎng)絡(luò)的高分影像場(chǎng)景分類[J];測(cè)繪學(xué)報(bào);2016年09期

3 楊格蘭;鄧曉軍;劉琮;;基于深度時(shí)空域卷積神經(jīng)網(wǎng)絡(luò)的表情識(shí)別模型[J];中南大學(xué)學(xué)報(bào)(自然科學(xué)版);2016年07期

4 劉帥師;程曦;郭文燕;陳奇;;深度學(xué)習(xí)方法研究新進(jìn)展[J];智能系統(tǒng)學(xué)報(bào);2016年05期

5 楊雨濃;房鼎益;王洪;;一種基于混合深度置信模型的面部表情識(shí)別方法[J];西南大學(xué)學(xué)報(bào)(自然科學(xué)版);2016年06期

6 陳耀丹;王連明;;基于卷積神經(jīng)網(wǎng)絡(luò)的人臉識(shí)別方法[J];東北師大學(xué)報(bào)(自然科學(xué)版);2016年02期

7 孫曉;潘汀;任福繼;;基于ROI-KNN卷積神經(jīng)網(wǎng)絡(luò)的面部表情識(shí)別[J];自動(dòng)化學(xué)報(bào);2016年06期

8 王偉凝;王勵(lì);趙明權(quán);蔡成加;師婷婷;徐向民;;基于并行深度卷積神經(jīng)網(wǎng)絡(luò)的圖像美感分類[J];自動(dòng)化學(xué)報(bào);2016年06期

9 馬曉;張番棟;封舉富;;基于深度學(xué)習(xí)特征的稀疏表示的人臉識(shí)別方法[J];智能系統(tǒng)學(xué)報(bào);2016年03期

10 牛連強(qiáng);陳向震;張勝男;王琪輝;;深度連續(xù)卷積神經(jīng)網(wǎng)絡(luò)模型構(gòu)建與性能分析[J];沈陽工業(yè)大學(xué)學(xué)報(bào);2016年06期

相關(guān)博士學(xué)位論文 前1條

1 萬川;基于動(dòng)態(tài)序列圖像的人臉表情識(shí)別系統(tǒng)理論與方法研究[D];吉林大學(xué);2013年

相關(guān)碩士學(xué)位論文 前10條

1 產(chǎn)文濤;基于卷積神經(jīng)網(wǎng)絡(luò)的人臉表情和性別識(shí)別[D];安徽大學(xué);2016年

2 劉曠;基于卷積網(wǎng)絡(luò)集成的面部表情識(shí)別方法[D];浙江大學(xué);2016年

3 陳向震;基于深度學(xué)習(xí)的人臉表情識(shí)別算法研究[D];沈陽工業(yè)大學(xué);2016年

4 曹寧;基于靜態(tài)圖像的人臉表情識(shí)別算法研究[D];西安科技大學(xué);2015年

5 池燕玲;基于深度學(xué)習(xí)的人臉識(shí)別方法的研究[D];福建師范大學(xué);2015年

6 施徐敢;基于深度學(xué)習(xí)的人臉表情識(shí)別[D];浙江理工大學(xué);2015年

7 汪濟(jì)民;基于卷積神經(jīng)網(wǎng)絡(luò)的人臉檢測(cè)和性別識(shí)別研究[D];南京理工大學(xué);2015年

8 葉浪;基于卷積神經(jīng)網(wǎng)絡(luò)的人臉識(shí)別研究[D];東南大學(xué);2015年

9 高旭;基于動(dòng)態(tài)序列圖像的人臉表情特征提取與識(shí)別[D];吉林大學(xué);2014年

10 劉銀華;LBP和深度信念網(wǎng)絡(luò)在非限制條件下人臉識(shí)別研究[D];五邑大學(xué);2014年

，

本文編號(hào)：2438608