基于卷積神經(jīng)網(wǎng)絡的場景分類的研究與應用
發(fā)布時間：2018-10-26 11:35
【摘要】：場景分類是圖像處理領域的重要研究方向之一。隨著計算機技術和互聯(lián)網(wǎng)的發(fā)展,大量的圖像數(shù)據(jù)涌入到人們的生活和工作中,面對如此巨大的圖像信息,傳統(tǒng)的場景分類方法和技術表現(xiàn)出很多不足。近年來,卷積神經(jīng)網(wǎng)絡(Convolutional Neural Network,CNN)在圖像處理領域取得了很多突破性進展,它是通過模擬人類大腦學習的過程,直接從圖像像素中提取圖像特征,并將特征提取與分類器結合到一個學習框架下,對相關對象進行分類識別。另外,卷積神經(jīng)網(wǎng)絡的局部連接、權值共享和降采樣大大減少了網(wǎng)絡的訓練參數(shù),簡化了網(wǎng)絡模型,進一步提高了網(wǎng)絡的訓練效率。本文針對場景圖像的復雜多變性和傳統(tǒng)場景分類方法泛化能力不強的問題,結合卷積神經(jīng)網(wǎng)絡方法進行場景分類。卷積神經(jīng)網(wǎng)絡分類性能的好壞主要決取于網(wǎng)絡的層次結構,因此本文研究了影響卷積神經(jīng)網(wǎng)絡分類性能的因素,并以此為根據(jù)設計了一個卷積神經(jīng)網(wǎng)絡模型,應用于場景分類中。具體工作如下:1.針對應用于場景分類設計的卷積神經(jīng)網(wǎng)絡模型中如何選擇層次結構問題,本文設計了一個淺層卷積神經(jīng)網(wǎng)絡模型,應用于Scene-15數(shù)據(jù)集和SUN-397數(shù)據(jù)集的場景圖像分類任務中,以此研究不同大小和個數(shù)的卷積核、不同的激活函數(shù)和不同采樣方法對卷積神經(jīng)網(wǎng)絡分類性能的影響。研究表明神經(jīng)網(wǎng)絡使用較小的卷積核以及較多的核數(shù)目、最大值采樣和使用ReLU激活函數(shù),可增加卷積神經(jīng)網(wǎng)絡的分類性能。2.為更好地適應實際場景圖像的要求,本文根據(jù)以上研究對神經(jīng)網(wǎng)絡模型進行了改進,設計了一個8層的卷積神經(jīng)網(wǎng)絡。該網(wǎng)絡的卷積層采用了較小的卷積核,并增加了卷積核的數(shù)量,這樣可以提取到更多的圖像特征,提高分類性能。同時,采樣層采用了最大值采樣方法以及ReLU激活函數(shù)。本文把改進后的卷積神經(jīng)網(wǎng)絡模型與AlexNet模型和VGGNet模型在Scene-15數(shù)據(jù)集和SUN-397數(shù)據(jù)集上進行了對比實驗,實驗結果證明了該模型在場景分類應用中具有良好的分類效果。本文主要是在MATLAB軟件上利用MatConvNet工具箱進行卷積神經(jīng)網(wǎng)絡的結構設計和參數(shù)優(yōu)化,分析了影響卷積神經(jīng)網(wǎng)絡分類性能的因素,并以此為根據(jù)設計了卷積神經(jīng)網(wǎng)絡模型,應用于場景分類中。大量實驗表明本文網(wǎng)絡模型在場景分類應用中具有良好的分類性能,并具有一定的泛化能力。
[Abstract]:Scene classification is one of the important research directions in the field of image processing. With the development of computer technology and the Internet, a large number of image data flow into people's lives and work. In the face of such huge image information, traditional scene classification methods and techniques show a lot of shortcomings. In recent years, convolutional neural network (Convolutional Neural Network,CNN) has made many breakthroughs in the field of image processing. It extracts image features directly from image pixels by simulating the learning process of human brain. The feature extraction and classifier are combined into a learning framework to classify and recognize the related objects. In addition, the local connection, weight sharing and down-sampling of convolutional neural networks greatly reduce the training parameters of the network, simplify the network model, and further improve the training efficiency of the network. Aiming at the complex variability of scene image and the weak generalization ability of traditional scene classification methods, this paper combines convolution neural network method to classify scene. The classification performance of convolutional neural networks is mainly determined by the hierarchical structure of the network. Therefore, the factors influencing the classification performance of convolutional neural networks are studied in this paper, based on which a convolution neural network model is designed. Applied to scene classification. The specific work is as follows: 1. Aiming at the problem of how to select hierarchical structure in the model of convolution neural network applied in scene classification design, a shallow convolution neural network model is designed in this paper, which is applied to the task of scene image classification in Scene-15 dataset and SUN-397 dataset. The effects of different size and number of convolution kernels, different activation functions and different sampling methods on the classification performance of convolution neural networks are studied. It is shown that the classification performance of convolutional neural networks can be improved by using smaller convolutional kernels and more kernel numbers, maximum sampling and ReLU activation function. 2. In order to better meet the requirements of the actual scene image, this paper improves the neural network model based on the above research, and designs an 8-layer convolution neural network. The convolution layer of the network uses smaller convolution cores and increases the number of convolution cores which can extract more image features and improve classification performance. At the same time, the maximum sampling method and the ReLU activation function are used in the sampling layer. In this paper, the improved convolution neural network model is compared with AlexNet model and VGGNet model on Scene-15 data set and SUN-397 data set. The experimental results show that the model has good classification effect in scene classification. In this paper, the structure design and parameter optimization of convolution neural network are carried out by using MatConvNet toolbox on MATLAB software. The factors influencing the classification performance of convolution neural network are analyzed, and the convolution neural network model is designed. Applied to scene classification. A large number of experiments show that the network model has good classification performance and generalization ability in scene classification.
【學位授予單位】：蘭州理工大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41;TP183
本文編號：2295584
[Abstract]:Scene classification is one of the important research directions in the field of image processing. With the development of computer technology and the Internet, a large number of image data flow into people's lives and work. In the face of such huge image information, traditional scene classification methods and techniques show a lot of shortcomings. In recent years, convolutional neural network (Convolutional Neural Network,CNN) has made many breakthroughs in the field of image processing. It extracts image features directly from image pixels by simulating the learning process of human brain. The feature extraction and classifier are combined into a learning framework to classify and recognize the related objects. In addition, the local connection, weight sharing and down-sampling of convolutional neural networks greatly reduce the training parameters of the network, simplify the network model, and further improve the training efficiency of the network. Aiming at the complex variability of scene image and the weak generalization ability of traditional scene classification methods, this paper combines convolution neural network method to classify scene. The classification performance of convolutional neural networks is mainly determined by the hierarchical structure of the network. Therefore, the factors influencing the classification performance of convolutional neural networks are studied in this paper, based on which a convolution neural network model is designed. Applied to scene classification. The specific work is as follows: 1. Aiming at the problem of how to select hierarchical structure in the model of convolution neural network applied in scene classification design, a shallow convolution neural network model is designed in this paper, which is applied to the task of scene image classification in Scene-15 dataset and SUN-397 dataset. The effects of different size and number of convolution kernels, different activation functions and different sampling methods on the classification performance of convolution neural networks are studied. It is shown that the classification performance of convolutional neural networks can be improved by using smaller convolutional kernels and more kernel numbers, maximum sampling and ReLU activation function. 2. In order to better meet the requirements of the actual scene image, this paper improves the neural network model based on the above research, and designs an 8-layer convolution neural network. The convolution layer of the network uses smaller convolution cores and increases the number of convolution cores which can extract more image features and improve classification performance. At the same time, the maximum sampling method and the ReLU activation function are used in the sampling layer. In this paper, the improved convolution neural network model is compared with AlexNet model and VGGNet model on Scene-15 data set and SUN-397 data set. The experimental results show that the model has good classification effect in scene classification. In this paper, the structure design and parameter optimization of convolution neural network are carried out by using MatConvNet toolbox on MATLAB software. The factors influencing the classification performance of convolution neural network are analyzed, and the convolution neural network model is designed. Applied to scene classification. A large number of experiments show that the network model has good classification performance and generalization ability in scene classification.
【學位授予單位】：蘭州理工大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41;TP183
【參考文獻】
相關期刊論文 前7條
1 顧廣華;韓晰瑛;陳春霞;趙耀;;圖像場景語義分類研究進展綜述[J];系統(tǒng)工程與電子技術;2016年04期
2 李學龍;史建華;董永生;陶大程;;場景圖像分類技術綜述[J];中國科學:信息科學;2015年07期
3 肖保良;;基于Gist特征與PHOG特征融合的多類場景分類[J];中北大學學報(自然科學版);2014年06期
4 趙理君;唐娉;霍連志;鄭柯;;圖像場景分類中視覺詞包模型方法綜述[J];中國圖象圖形學報;2014年03期
5 ZHOU Li;HU DeWen;ZHOU ZongTan;;Scene recognition combining structural and textural features[J];Science China(Information Sciences);2013年07期
6 楊昭;高雋;謝昭;吳克偉;;局部Gist特征匹配核的場景分類[J];中國圖象圖形學報;2013年03期
7 吳偉仁;王大軼;邢琰;龔小謹;劉濟林;;月球車巡視探測的雙目視覺里程算法與實驗研究[J];中國科學:信息科學;2011年12期
相關碩士學位論文 前4條
1 吳正文;卷積神經(jīng)網(wǎng)絡在圖像分類中的應用研究[D];電子科技大學;2015年
2 劉靜;基于特征組合的圖像場景分類[D];湘潭大學;2015年
3 周圣云;基于視覺感知的室內場景識別與理解[D];電子科技大學;2015年
4 李飛騰;卷積神經(jīng)網(wǎng)絡及其應用[D];大連理工大學;2014年
，本文編號：2295584
本文鏈接：http://sikaile.net/kejilunwen/zidonghuakongzhilunwen/2295584.html
最近更新
教材專著
