基于深度學習的圖像識別方法研究與應用
發(fā)布時間：2019-05-09 14:59
【摘要】：圖像識別是圖像研究領(lǐng)域中的一個重要研究方向,也是機器視覺中的熱點研究問題,具有非常重大的意義。深度學習,近些年在圖像、語音、文本等方面取得了許多成果。同時,深度學習在人工智能領(lǐng)域占據(jù)著重要的地位,在日常生活中受到廣泛的應用和關(guān)注。傳統(tǒng)的圖像識別方法需要人工設計特征,相對依賴圖像識別經(jīng)驗豐富的研究學者,且傳統(tǒng)的方法圖像識別率較低。隨著互聯(lián)網(wǎng)和信息技術(shù)的發(fā)展,大數(shù)據(jù)背景下產(chǎn)生的海量圖像數(shù)據(jù),傳統(tǒng)的識別方法已經(jīng)不能滿足我們的需求。而深度學習是一個多層的網(wǎng)絡結(jié)構(gòu),通過模擬人腦,能夠自動的學習和提取特征,充分發(fā)揮大數(shù)據(jù)的優(yōu)勢。因此,本文將深度學習和圖像識別相結(jié)合,研究如何提高圖像的識別率,具有一定的研究空間和研究價值。本文首先闡述了圖像識別和深度學習的理論,與淺層學習相比,深度學習能夠容易的表達復雜函數(shù),具有很強的泛化能力。同時,還探討了幾種常用的深度學習模型及其算法原理,研究了圖像的特征提取和識別方法。本文在研究深度神經(jīng)網(wǎng)絡的基礎(chǔ)上,針對原始的初始化權(quán)重方法造成的網(wǎng)絡學習速度慢的問題,提出了改進的初始化權(quán)重方法。同時,在理論和實驗上驗證了該方法的有效性,還可以將其運用到常用的卷積神經(jīng)網(wǎng)絡和深度信念網(wǎng)絡中。其次,由于深度神經(jīng)網(wǎng)絡存在梯度消失的問題。同時,深度信念網(wǎng)絡的半監(jiān)督學習特點,可以挖掘大量無標簽數(shù)據(jù)的價值。因此,本論文提出了改進的深度信念網(wǎng)絡學習模型。通過實驗證明,該模型的學習速度和識別正確率都得到提高。相對于未改進的深度信念網(wǎng)絡,該模型在MNIST數(shù)據(jù)集上的識別率達到了99.18%,提高了 0.62%,在CIFAR-10數(shù)據(jù)集上的識別率提高了 9.6%。最后,針對卷積神經(jīng)網(wǎng)絡特別適合處理與圖像相關(guān)的問題,本文提出了改進的卷積神經(jīng)網(wǎng)絡模型。該模型首先采用改進的初始化權(quán)重方法代替原始的初始化方法;然后去掉池化層,采用SVM分類器替代了原始的softmax層;最后對激活函數(shù)進行了改進,改進后的函數(shù)結(jié)合了 Sigmoid函數(shù)的光滑性和ReLU函數(shù)的稀疏性及快速收斂性等特點,同時引入了 Dropout思想,目的是為了增強網(wǎng)絡泛化的能力,防止網(wǎng)絡過擬合。該模型在MNIST數(shù)據(jù)集上的識別率達到了 99.52%,相對于未改進的卷積神經(jīng)網(wǎng)絡,提高了 0.66%,與傳統(tǒng)方法相比,提高了 5%左右。在CIFAR-10數(shù)據(jù)集上,與未改進的卷積神經(jīng)網(wǎng)絡相比,識別正確率提高了 6.4%,與傳統(tǒng)方法相比,提高了 9%左右。通過實驗表明,該模型的有效性得到驗證,表現(xiàn)效果較好,圖像的識別率得到提高。
[Abstract]:Image recognition is an important research direction in the field of image research, and it is also a hot research topic in machine vision, which is of great significance. In recent years, in-depth learning has made a lot of achievements in image, voice, text and so on. At the same time, deep learning occupies an important position in the field of artificial intelligence and has been widely used and concerned in daily life. The traditional image recognition method needs manual design features, which depends on the experienced researchers of image recognition, and the image recognition rate of the traditional method is low. With the development of Internet and information technology, the traditional recognition methods can not meet our needs for the massive image data produced under the background of big data. Deep learning is a multi-layer network structure, which can automatically learn and extract features and give full play to big data's advantages by simulating the human brain. Therefore, this paper combines depth learning with image recognition to study how to improve the recognition rate of images, which has a certain research space and research value. In this paper, the theory of image recognition and deep learning is expounded. Compared with shallow learning, deep learning can express complex functions easily and has strong generalization ability. At the same time, several kinds of depth learning models and their algorithm principles are discussed, and the feature extraction and recognition methods of images are studied. In this paper, based on the study of deep neural network, an improved initialization weight method is proposed to solve the problem of slow network learning speed caused by the original initialization weight method. At the same time, the effectiveness of the method is verified theoretically and experimentally, and it can also be applied to convolution neural networks and deep belief networks. Secondly, the depth neural network has the problem of gradient disappearance. At the same time, the semi-supervised learning characteristics of the deep belief network can mine the value of a large amount of untagged data. Therefore, this paper proposes an improved in-depth belief network learning model. The experimental results show that the learning speed and recognition accuracy of the model are improved. Compared with the unimproved deep belief network, the recognition rate of the model on MNIST dataset is 99.18%, increased by 0.62%, and the recognition rate on CIFAR-10 dataset is increased by 9.6%. Finally, an improved convolution neural network model is proposed to deal with image-related problems. In this model, the improved initialization weight method is used to replace the original initialization method, and then the pooling layer is removed, and the SVM classifier is used to replace the original softmax layer. Finally, the activation function is improved. the improved function combines the smoothness of Sigmoid function and the sparsity and fast convergence of ReLU function, and introduces the idea of Dropout in order to enhance the ability of network generalization. Prevent network overfitting. The recognition rate of the model on MNIST dataset is 99.52%. Compared with the unimproved convolution neural network, the recognition rate of the model is increased by 0.66%, which is about 5% higher than that of the traditional method. On CIFAR-10 datasets, compared with the unimproved convolution neural network, the recognition accuracy is improved by 6.4% and by about 9% compared with the traditional method. The experimental results show that the effectiveness of the model is verified, the performance is better and the recognition rate of the image is improved.
【學位授予單位】：華中師范大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41
本文編號：2472855
[Abstract]:Image recognition is an important research direction in the field of image research, and it is also a hot research topic in machine vision, which is of great significance. In recent years, in-depth learning has made a lot of achievements in image, voice, text and so on. At the same time, deep learning occupies an important position in the field of artificial intelligence and has been widely used and concerned in daily life. The traditional image recognition method needs manual design features, which depends on the experienced researchers of image recognition, and the image recognition rate of the traditional method is low. With the development of Internet and information technology, the traditional recognition methods can not meet our needs for the massive image data produced under the background of big data. Deep learning is a multi-layer network structure, which can automatically learn and extract features and give full play to big data's advantages by simulating the human brain. Therefore, this paper combines depth learning with image recognition to study how to improve the recognition rate of images, which has a certain research space and research value. In this paper, the theory of image recognition and deep learning is expounded. Compared with shallow learning, deep learning can express complex functions easily and has strong generalization ability. At the same time, several kinds of depth learning models and their algorithm principles are discussed, and the feature extraction and recognition methods of images are studied. In this paper, based on the study of deep neural network, an improved initialization weight method is proposed to solve the problem of slow network learning speed caused by the original initialization weight method. At the same time, the effectiveness of the method is verified theoretically and experimentally, and it can also be applied to convolution neural networks and deep belief networks. Secondly, the depth neural network has the problem of gradient disappearance. At the same time, the semi-supervised learning characteristics of the deep belief network can mine the value of a large amount of untagged data. Therefore, this paper proposes an improved in-depth belief network learning model. The experimental results show that the learning speed and recognition accuracy of the model are improved. Compared with the unimproved deep belief network, the recognition rate of the model on MNIST dataset is 99.18%, increased by 0.62%, and the recognition rate on CIFAR-10 dataset is increased by 9.6%. Finally, an improved convolution neural network model is proposed to deal with image-related problems. In this model, the improved initialization weight method is used to replace the original initialization method, and then the pooling layer is removed, and the SVM classifier is used to replace the original softmax layer. Finally, the activation function is improved. the improved function combines the smoothness of Sigmoid function and the sparsity and fast convergence of ReLU function, and introduces the idea of Dropout in order to enhance the ability of network generalization. Prevent network overfitting. The recognition rate of the model on MNIST dataset is 99.52%. Compared with the unimproved convolution neural network, the recognition rate of the model is increased by 0.66%, which is about 5% higher than that of the traditional method. On CIFAR-10 datasets, compared with the unimproved convolution neural network, the recognition accuracy is improved by 6.4% and by about 9% compared with the traditional method. The experimental results show that the effectiveness of the model is verified, the performance is better and the recognition rate of the image is improved.
【學位授予單位】：華中師范大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41
【參考文獻】
相關(guān)期刊論文 前10條
1 李丹;沈夏炯;張海香;朱永強;;基于Lenet-5的卷積神經(jīng)網(wǎng)絡改進算法[J];計算機時代;2016年08期
2 朱常寶;程勇;高強;;基于半監(jiān)督深度信念網(wǎng)絡的圖像分類算法研究[J];計算機科學;2016年S1期
3 張陽;劉偉銘;吳義虎;;基于深信度網(wǎng)絡分類算法的行人檢測方法[J];計算機應用研究;2016年02期
4 張建明;詹智財;成科揚;詹永照;;深度學習的研究與發(fā)展[J];江蘇大學學報(自然科學版);2015年02期
5 楊陽;張文生;;基于深度學習的圖像自動標注算法[J];數(shù)據(jù)采集與處理;2015年01期
6 尹寶才;王文通;王立春;;深度學習研究綜述[J];北京工業(yè)大學學報;2015年01期
7 楊俊安;王一;劉輝;李晉徽;陸俊;;深度學習理論及其在語音識別領(lǐng)域的應用[J];通信對抗;2014年03期
8 黃凱奇;任偉強;譚鐵牛;;圖像物體分類與檢測算法綜述[J];計算機學報;2014年06期
9 鄭胤;陳權(quán)崎;章毓晉;;深度學習及其在目標和行為識別中的新進展[J];中國圖象圖形學報;2014年02期
10 李星;郭曉松;郭君斌;;基于HOG特征和SVM的前向車輛識別方法[J];計算機科學;2013年S2期
相關(guān)碩士學位論文 前10條
1 張飛云;基于深度學習的車輛定位及車型識別研究[D];江蘇大學;2016年
2 李卓;基于深度學習的字符識別系統(tǒng)的設計與實現(xiàn)[D];北京林業(yè)大學;2016年
3 張超群;基于深度學習的字符識別[D];電子科技大學;2016年
4 楊楠;基于Caffe深度學習框架的卷積神經(jīng)網(wǎng)絡研究[D];河北師范大學;2016年
5 楊東坡;基于深度學習的商品圖像分類[D];大連交通大學;2015年
6 楚敏南;基于卷積神經(jīng)網(wǎng)絡的圖像分類技術(shù)研究[D];湘潭大學;2015年
7 李曉普;基于卷積神經(jīng)網(wǎng)絡的圖像分類[D];大連理工大學;2015年
8 謝寶劍;基于卷積神經(jīng)網(wǎng)絡的圖像分類方法研究[D];合肥工業(yè)大學;2015年
9 白亞龍;基于深度神經(jīng)網(wǎng)絡的圖像識別系統(tǒng)的研究與改進[D];哈爾濱工業(yè)大學;2014年
10 楊心;基于卷積神經(jīng)網(wǎng)絡的交通標識識別研究與應用[D];大連理工大學;2014年
，本文編號：2472855
本文鏈接：http://sikaile.net/kejilunwen/ruanjiangongchenglunwen/2472855.html
最近更新
教材專著
