本文關(guān)鍵詞： 改進LBP 卷積神經(jīng)網(wǎng)絡(luò) 深度信念網(wǎng) 支持向量機 超限學(xué)習(xí)機　出處：《青島大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

【摘要】：B超圖像中肝硬化的識別在臨床上對肝病的診斷具有重要意義,但是臨床采集的圖像中有噪聲、尺度不一、病變邊緣模糊、回聲不均勻等因素。傳統(tǒng)的圖像識別技術(shù)在對肝硬化進行識別時,準確率較低,主要是由于傳統(tǒng)的特征提取算法不能很好地描述肝硬化病變的特點。針對上述問題,本文提出了一種改進的局部二值模式的特征提取算法,很好地解決了病變區(qū)域邊緣模糊的問題;利用深度學(xué)習(xí)理論,設(shè)計了兩種卷積神經(jīng)網(wǎng)絡(luò)模型和一種深度信念網(wǎng)絡(luò)模型,得到了表現(xiàn)力更強的特征。本文主要工作有以下幾個方面:(1)將局部二值模式(Local Binary Pattern,LBP)算法進行改進,并通過信息熵來度量通過該改進LBP算法得到的樣本的信息量。選出包含紋理信息最多的樣本,作為支持向量機(Support Vector Machine,SVM)和超限學(xué)習(xí)機(Extreme Learning Machine,ELM)的輸入。實驗結(jié)果表明,所提出的改進算法對于提高肝硬化的識別率很有幫助。(2)利用深度學(xué)習(xí)思想,設(shè)計了兩種基于卷積神經(jīng)網(wǎng)絡(luò)(Convolutional Neural Networks)的模型,分別是Liver-CNN1和Liver-CNN2,并將其與SVM和ELM結(jié)合進行分類,替換了普通卷積神經(jīng)網(wǎng)絡(luò)最后的全連接神經(jīng)網(wǎng)絡(luò),不僅取得了更高的分類準確率,而且在分類時間上也有明顯的減少,具有實時性。卷積神經(jīng)網(wǎng)絡(luò)中感受野概念和權(quán)值共享技術(shù)的使用,大大減少了網(wǎng)絡(luò)參數(shù)的數(shù)量,提高了學(xué)習(xí)的速度。經(jīng)實驗對比得知,卷積神經(jīng)網(wǎng)絡(luò)自學(xué)習(xí)得到的特征相對于傳統(tǒng)的特征有更高的識別率,說明本文的兩種卷積神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)到的特征對肝硬化病變有更準確的描述。(3)利用深度信念網(wǎng)(Deep Belief Nets,DBNs)與SVM和ELM結(jié)合的分類方法,將一般的深度信念網(wǎng)的最頂層的邏輯回歸,用SVM和ELM代替,省去了回歸值與類別之間的度量過程。而且利用DBNs的預(yù)訓(xùn)練和分層學(xué)習(xí),得到了更接近全局最優(yōu)的解,網(wǎng)絡(luò)收斂更快。
[Abstract]:The recognition of liver cirrhosis in B-ultrasound images is of great significance in the diagnosis of liver diseases. However, the images collected in the clinic have noise, different scales and blurred lesion margins. Traditional image recognition technology has low accuracy in the recognition of liver cirrhosis. The main reason is that the traditional feature extraction algorithm can not describe the characteristics of liver cirrhosis well. In view of the above problems, this paper proposes an improved feature extraction algorithm of local binary pattern. The fuzzy edge of the lesion area is well solved. Two kinds of convolution neural network models and a deep belief network model are designed by using the theory of deep learning. The main work of this paper is as follows: 1) improve the local binary mode (Local Binary pattern) algorithm. The information entropy is used to measure the information of the samples obtained by the improved LBP algorithm, and the samples with the most texture information are selected. As support Vector Machine (SVM) and extreme Learning Machine. The experimental results show that the improved algorithm is helpful to improve the recognition rate of liver cirrhosis. Two models based on Convolutional Neural Networks are designed. Liver-CNN1 and Liver-CNN2 were combined with SVM and ELM to replace the final fully connected neural network of the general convolutional neural network. It not only achieves higher classification accuracy, but also has a significant reduction in classification time, real-time. The concept of perception field and weight sharing technology are used in convolution neural network. It greatly reduces the number of network parameters and improves the speed of learning. The experimental results show that the characteristics obtained by convolution neural networks have a higher recognition rate than the traditional features. It is suggested that the features of the two convolutional neural networks in this paper are more accurate in describing the liver cirrhosis lesions.) Deep Belief Nets is used in this paper. DBNs) combined with SVM and ELM, the logical regression of the top layer of general deep belief net is replaced by SVM and ELM. The measurement process between regression value and category is omitted, and by using the pretraining and hierarchical learning of DBNs, the global optimal solution is obtained, and the convergence of the network is faster.
【學(xué)位授予單位】：青島大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：R575.2;TP18;TP391.41

【參考文獻】

相關(guān)期刊論文 前9條

1 雷一鳴;趙希梅;于可歆;王國棟;郭衛(wèi)東;;魯棒的超聲圖像肝硬化識別方法[J];電子測量與儀器學(xué)報;2016年10期

2 張慧;遲慶云;劉彩霞;;基于灰度共生矩陣的肝癌B超紋理特征決策樹診斷分析[J];中國醫(yī)藥指南;2015年25期

3 程文博;張云;周華民;崔樹標;;基于卷積神經(jīng)網(wǎng)絡(luò)的注塑制品短射缺陷識別[J];塑料工業(yè);2015年07期

4 尹寶才;王文通;王立春;;深度學(xué)習(xí)研究綜述[J];北京工業(yè)大學(xué)學(xué)報;2015年01期

5 趙志剛;吳鑫;洪丹楓;潘振寬;;基于信息熵的GLBP掌紋識別算法[J];計算機科學(xué);2014年08期

6 潘聞特;申麗萍;;基于BM神經(jīng)網(wǎng)絡(luò)編碼的生理信號情感識別[J];計算機工程與設(shè)計;2012年03期

7 吳證;周越;杜春華;袁泉;;結(jié)合主元成分分析的受限玻耳茲曼機神經(jīng)網(wǎng)絡(luò)的降維方法[J];上海交通大學(xué)學(xué)報;2008年04期

8 陳菲;;基于BP神經(jīng)網(wǎng)絡(luò)的超聲肝圖像識別[J];兵工自動化;2007年08期

9 周國輝,汪源源,王威琪,孫英,陳悅;基于超聲圖像特征的肝硬化分析系統(tǒng)[J];聲學(xué)技術(shù);2003年04期

相關(guān)博士學(xué)位論文 前1條

1 羅恒;基于協(xié)同過濾視角的受限玻爾茲曼機研究[D];上海交通大學(xué);2011年

相關(guān)碩士學(xué)位論文 前3條

1 汪海波;基于深度學(xué)習(xí)的人臉識別方法研究[D];合肥工業(yè)大學(xué);2014年

2 李衛(wèi);深度學(xué)習(xí)在圖像識別中的研究及應(yīng)用[D];武漢理工大學(xué);2014年

3 林妙真;基于深度學(xué)習(xí)的人臉識別研究[D];大連理工大學(xué);2013年

，

本文編號：1445335