本文關(guān)鍵詞：基于深度學(xué)習(xí)的圖像分類方法研究　出處：《中國礦業(yè)大學(xué)》2016年碩士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 圖像分類 卷積神經(jīng)網(wǎng)絡(luò) 深度置信網(wǎng) 自動(dòng)編碼器 極速學(xué)習(xí)

【摘要】：圖像作為人類感知事物的視覺基礎(chǔ),是人們從外界獲得信息的重要依據(jù),所以讓機(jī)器自動(dòng)完成圖像識(shí)別、分類具有重要意義。圖像分類最重要的部分是特征提取,研究高效的特征提取算法在圖像領(lǐng)域至關(guān)重要。深度學(xué)習(xí)(Deep Learning, DL)是多層的網(wǎng)絡(luò)結(jié)構(gòu),它通過建立、模擬人腦的分層結(jié)構(gòu),對(duì)外部輸入的聲音、圖像、文本等數(shù)據(jù)進(jìn)行從低級(jí)到高級(jí)的特征提取,所以深度學(xué)習(xí)在圖像分類領(lǐng)域具有廣闊的應(yīng)用空間。而深度學(xué)習(xí)本身存在訓(xùn)練時(shí)間過長(zhǎng)、過擬合等問題,本文以提高深度模型分類精確度、縮短訓(xùn)練時(shí)間和防止模型過擬合三個(gè)問題為出發(fā)點(diǎn),主要研究工作如下：首先,本文研究了極速學(xué)習(xí)機(jī)(Extreme Learning Machine,ELM)作為卷積神經(jīng)網(wǎng)絡(luò)(Convolutional Neural Network,CNN)分類器的可行性與意義,進(jìn)而提出了混合深度模型CNN-ELM(Convolutional Neural Network-Extreme Learning Machine)。先用原始的CNN訓(xùn)練網(wǎng)絡(luò),然后用ELM替換CNN的輸出層完成最后的分類,混合模型結(jié)合了CNN有效提取圖像特征和ELM快速高效的特點(diǎn),使得兩種方法能夠協(xié)同工作,實(shí)驗(yàn)表明CNN-ELM提高了CNN的分類精確度。其次,針對(duì)深度學(xué)習(xí)方法訓(xùn)練時(shí)間過長(zhǎng)的問題,研究了隨機(jī)參數(shù)網(wǎng)絡(luò)結(jié)構(gòu)的可行性與意義。核極速學(xué)習(xí)機(jī)是在ELM的基礎(chǔ)上引入了核函數(shù),具有更好的分類效果,從而提出了基于核極速學(xué)習(xí)機(jī)的隨機(jī)參數(shù)深度模型：卷積極速學(xué)習(xí)機(jī)(Convolutional Extreme Learning Machine with Kernel,CKELM)。在模型CKELM中,把隨機(jī)權(quán)值的卷積層和降采樣層作為隱含層,來提取輸入圖像的顯著特征。實(shí)驗(yàn)表明該算法既保證了分類精確度又大大縮短了深度算法的訓(xùn)練時(shí)間。最后,本文研究了基于DropConnect的深度自動(dòng)編碼器算法的應(yīng)用意義和可行性。DropConnect作為一種新型的正則化方法,在處理過擬合等問題上表現(xiàn)突出,所以文章提出了一種基于DropConnect的深度自動(dòng)編碼器模型DDAE(DropConnect Deep AutoEncoder)。實(shí)驗(yàn)表明將DropConnect思想引入自動(dòng)編碼器中有效的提高了算法的性能。
[Abstract]:As the visual basis of human perception, image is an important basis for people to obtain information from the outside world, so the machine can automatically complete image recognition. Classification is of great significance. The most important part of image classification is feature extraction. It is very important to study efficient feature extraction algorithm in the field of image. DL) is a multi-layer network structure, it builds, simulates the human brain's hierarchical structure, carries on the low-level to the high-level feature extraction to the external input sound, the image, the text and so on data. So depth learning has a wide application space in the field of image classification, and depth learning itself has the problems of too long training time and over-fitting, so this paper improves the accuracy of depth model classification. Shortening the training time and preventing the model from overfitting is the starting point. The main research work is as follows: first. This paper studies extreme Learning Machine. The feasibility and significance of ELM as Convolutional Neural Network classifier. Furthermore, the mixed depth model CNN-ELM (. Convolutional Neural Network-Extreme Learning Machine. First use the original CNN to train the network. Then the final classification is completed by replacing the output layer of CNN with ELM. The hybrid model combines the features of CNN extraction and the fast and efficient feature of ELM, which makes the two methods work together. Experiments show that CNN-ELM improves the classification accuracy of CNN. Secondly, the training time of deep learning method is too long. The feasibility and significance of random parameter network structure are studied. The kernel function is introduced into the kernel pole learning machine based on ELM, which has better classification effect. Thus a random parameter depth model based on kernel pole learning machine is proposed: convolution extreme speed learning machine (. Convolutional Extreme Learning Machine with Kernel. In model CKELM, the convolution layer and downsampling layer of random weights are used as hidden layers. Experiments show that the algorithm not only ensures the classification accuracy but also greatly reduces the training time of the depth algorithm. This paper studies the application significance and feasibility of depth automatic encoder algorithm based on DropConnect. DropConnect is a new regularization method. In dealing with problems such as fitting outstanding performance. In this paper, a depth automatic encoder model DDAE(DropConnect Deep AutoEncoder based on DropConnect is proposed. The experimental results show that the performance of the algorithm is improved effectively by introducing the DropConnect idea into the automatic encoder.
【學(xué)位授予單位】：中國礦業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.41

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 宋志堅(jiān);余銳;;基于深度學(xué)習(xí)的手寫數(shù)字分類問題研究[J];重慶工商大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年08期

2 郭麗麗;丁世飛;;深度學(xué)習(xí)研究進(jìn)展[J];計(jì)算機(jī)科學(xué);2015年05期

3 劉勘;袁蘊(yùn)英;;基于自動(dòng)編碼器的短文本特征提取及聚類研究[J];北京大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年02期

4 曲建嶺;杜辰飛;邸亞洲;高峰;郭超然;;深度自動(dòng)編碼器的研究與展望[J];計(jì)算機(jī)與現(xiàn)代化;2014年08期

5 王勇;趙儉輝;章登義;葉威;;基于稀疏自編碼深度神經(jīng)網(wǎng)絡(luò)的林火圖像分類[J];計(jì)算機(jī)工程與應(yīng)用;2014年24期

6 申豐山;王黎明;張軍英;;基于SVM技術(shù)的精簡(jiǎn)極速學(xué)習(xí)機(jī)[J];華中科技大學(xué)學(xué)報(bào)(自然科學(xué)版);2014年06期

7 王杰;畢浩洋;;基于正則極限學(xué)習(xí)機(jī)的煙草病毒病預(yù)測(cè)[J];鄭州大學(xué)學(xué)報(bào)(理學(xué)版);2013年04期

8 王杰;郭晨龍;;小波核極限學(xué)習(xí)機(jī)分類器[J];微電子學(xué)與計(jì)算機(jī);2013年10期

9 趙元慶;吳華;;多尺度特征和神經(jīng)網(wǎng)絡(luò)相融合的手寫體數(shù)字識(shí)別[J];計(jì)算機(jī)科學(xué);2013年08期

10 徐彩云;;圖像識(shí)別技術(shù)研究綜述[J];電腦知識(shí)與技術(shù);2013年10期

，

本文編號(hào)：1367434