本文選題：卷積神經(jīng)網(wǎng)絡(luò) + 圖像分類　； 參考：《濟南大學》2017年碩士論文

【摘要】：隨著經(jīng)濟的發(fā)展和社會的高速進步,圖像數(shù)據(jù)在我們?nèi)粘Ｉ钪邪l(fā)揮著越來越重要的作用。圖像數(shù)據(jù)的爆炸式增長使得需要分類的事物種類越來越多,而且被分類的對象內(nèi)容也越來越復(fù)雜。傳統(tǒng)的圖像分類方法已經(jīng)不能滿足現(xiàn)實應(yīng)用的需要,如何在大數(shù)據(jù)下提高圖像分類的準確率意義重大。卷積神經(jīng)網(wǎng)絡(luò)(Convolutional Neural Network,CNN)是一個新型的人工神經(jīng)網(wǎng)絡(luò)方法,在處理二維圖像領(lǐng)域表現(xiàn)出良好的性能,因此卷積神經(jīng)網(wǎng)絡(luò)被廣泛地應(yīng)用在圖像分類領(lǐng)域。圖像分類的正確率受卷積神經(jīng)網(wǎng)絡(luò)結(jié)構(gòu)的影響,因此研究卷積神經(jīng)網(wǎng)絡(luò)結(jié)構(gòu)優(yōu)化問題具有重要的理論價值和實用價值。本文分析了卷積神經(jīng)網(wǎng)絡(luò)的基本概念和算法,在經(jīng)典的卷積神經(jīng)網(wǎng)絡(luò)基礎(chǔ)上,主要進行了以下兩方面的工作:(1)基于經(jīng)典的PCA network(PCANET)網(wǎng)絡(luò)結(jié)構(gòu),在非線性激活函數(shù)之前引入maxout神經(jīng)網(wǎng)絡(luò),用softmax分類器替換SVM分類器,構(gòu)建了PCA非監(jiān)督預(yù)訓練的maxout卷積神經(jīng)網(wǎng)絡(luò)。該網(wǎng)絡(luò)參數(shù)求解過程中不需要調(diào)參技巧,訓練時間短,卷積核求解不需要反復(fù)迭代,且能適應(yīng)不同的圖像分類任務(wù)。網(wǎng)絡(luò)的整體流程分為五個階段:第一個階段:PCA非監(jiān)督預(yù)訓練學習濾波器,學習到的濾波器與圖片進行卷積提取圖像的特征;第二個階段:提取的特征經(jīng)過maxout神經(jīng)網(wǎng)絡(luò)后再輸入到非線性激活函數(shù)Relu中;第三個階段:非線性激活函數(shù)的輸出進行二值化,得到新的特征圖;第四個階段:新的特征圖分塊直方圖統(tǒng)計,列向量化輸入全連接層中;第五個階段:利用softmax分類器進行分類。在手寫體MNIST及其變形數(shù)據(jù)庫和自然圖像CIFAR-10數(shù)據(jù)庫上的實驗結(jié)果表明,PCA非監(jiān)督預(yù)訓練的maxout卷積神經(jīng)網(wǎng)絡(luò)的分類準確率有一定程度的提高。(2)基于經(jīng)典的Network in Network(NIN)網(wǎng)絡(luò)結(jié)構(gòu),對輸入圖像像素重構(gòu),構(gòu)建了基于雙邊濾波的多路徑卷積神經(jīng)網(wǎng)絡(luò)。該網(wǎng)絡(luò)減少了復(fù)雜圖像特征提取過程中前景物體紋理和形狀信息的丟失。網(wǎng)絡(luò)的輸入為兩個路徑,一個路徑輸入原始的圖像,另一個路徑輸入原始圖像像素重構(gòu)之后的圖像,兩個路徑獨立地提取特征,最后在均值降采樣層之后將兩個路徑提取的特征向量進行合并,輸入softmax分類器中進行分類。在自然圖像CIFAR-100數(shù)據(jù)庫上,分析圖像的復(fù)雜性和卷積神經(jīng)網(wǎng)絡(luò)在不同復(fù)雜度圖像上的學習曲線,得出卷積層和降采樣層提取的特征向量中前景物體紋理和形狀信息的丟失導致復(fù)雜度高的圖像易被錯誤分類。在自然圖像CIFAR-10和CIFAR-100數(shù)據(jù)庫上,通過實驗驗證了基于雙邊濾波的多路徑卷積神經(jīng)網(wǎng)絡(luò)取得的圖像分類準確率優(yōu)于傳統(tǒng)的單路徑卷積神經(jīng)網(wǎng)絡(luò)。
[Abstract]:With the development of economy and the rapid progress of society, image data plays a more and more important role in our daily life. The explosive growth of image data makes more and more kinds of things need to be classified, and the contents of objects are becoming more and more complex. Traditional image classification methods can not meet the needs of practical applications. How to improve the accuracy of image classification under big data is of great significance. Convolutional neural network (CNN) is a new artificial neural network method, which has good performance in two-dimensional image processing. Therefore, convolutional neural network is widely used in image classification. The accuracy of image classification is influenced by the network structure of convolution neural network, so it is of great theoretical and practical value to study the optimization of convolution neural network structure. In this paper, the basic concepts and algorithms of convolution neural network are analyzed. On the basis of classical convolution neural network, the following two main works are carried out: (1) based on the classical PCA network (PCANET) network structure, The maxout neural network is introduced before the nonlinear activation function and the softmax classifier is replaced by the softmax classifier to construct the unsupervised maxout convolution neural network. The parameters of the network do not need parameter adjustment technique, the training time is short, the convolution kernel solution does not need repeated iterations, and it can adapt to different image classification tasks. The whole flow of the network is divided into five stages: the first stage is the unsupervised pretraining filter of the PCA, which extracts the feature of the image by convolution between the filter and the picture. In the second stage, the extracted features are input into the nonlinear activation function Relu after the maxout neural network, the third stage: the output of the nonlinear activation function is binary, and the new feature diagram is obtained. The fourth stage: new feature graph block histogram statistics, column quantization input into the full join layer, the fifth stage: the use of softmax classifier to classify. The experimental results on handwritten MNIST and its deformation database and natural image CIFAR-10 database show that the classification accuracy of unsupervised maxout convolution neural network is improved to some extent. (2) based on the classical Network in Network (NIN) network structure, the classification accuracy of maxout convolution neural network is improved to some extent. The multipath convolution neural network based on bilateral filtering is constructed for pixel reconstruction of input image. The network reduces the loss of texture and shape information of foreground objects in the process of feature extraction of complex images. The input of the network is two paths, one path inputs the original image, the other path inputs the original image pixel reconstructed image, and the two paths extract the feature independently. Finally, the feature vectors extracted from the two paths are merged after the average downsampling layer, and the feature vectors are input into the softmax classifier for classification. In the CIFAR-100 database of natural images, the complexity of images and the learning curves of convolution neural networks on different complexity images are analyzed. It is concluded that the loss of texture and shape information of foreground objects in the feature vectors extracted by convolution layer and de-sampling layer leads to high complexity images being easily misclassified. On the CIFAR-10 and CIFAR-100 databases of natural images, the experimental results show that the classification accuracy of the multi-path convolution neural network based on bilateral filtering is better than that of the traditional single-path convolution neural network.
【學位授予單位】：濟南大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41;TP183

【參考文獻】

相關(guān)期刊論文 前5條

1 王志瑞;閆彩良;;圖像特征提取方法的綜述[J];吉首大學學報(自然科學版);2011年05期

2 章志勇,潘志庚,張明敏,吳海虹;基于多尺度通用傅里葉描述子的灰度圖像檢索[J];中國圖象圖形學報;2005年05期

3 李智勇,匡綱要,鄒煥新;投影追蹤方法在高光譜圖像異常點檢測中的應(yīng)用[J];電子與信息學報;2004年09期

4 楊健,楊靜宇,葉暉;Fisher線性鑒別分析的理論研究及其應(yīng)用[J];自動化學報;2003年04期

5 董火明,高雋,胡良梅,董文雯;基于主分量分析的形狀特征提取及識別研究[J];合肥工業(yè)大學學報(自然科學版);2003年02期

相關(guān)碩士學位論文 前1條

1 劉三軍;基于圖像特征空間學習的圖像分類方法研究[D];西安電子科技大學;2014年

，

本文編號：2113458