發(fā)布時(shí)間：2018-04-01 06:37

  本文選題：深度學(xué)習(xí)　切入點(diǎn)：自編碼機(jī)　出處：《武漢大學(xué)》2017年碩士論文

【摘要】：近年來,隨著成像技術(shù)的發(fā)展,圖像、視頻等數(shù)據(jù)越來越豐富,如何從海量數(shù)據(jù)中提取有效信息也成為了一個(gè)難題。深度學(xué)習(xí)已被證明是有效的解決途徑。深度神經(jīng)網(wǎng)絡(luò)是一種多層次特征學(xué)習(xí)模型,能夠自動(dòng)從原始視覺數(shù)據(jù)中提取出較為抽象的特征,用于后續(xù)的圖像分類、目標(biāo)檢測等工作。然而,深度神經(jīng)網(wǎng)絡(luò)的訓(xùn)練往往需要大量的標(biāo)記樣本,并且模型參數(shù)數(shù)量龐大,容易過擬合。為了解決這些問題,本文基于前人的模型,提出了改進(jìn)的深度神經(jīng)網(wǎng)絡(luò)模型以及防止過擬合的方法,以提高深度神經(jīng)網(wǎng)絡(luò)在圖像識(shí)別中的性能。本文的創(chuàng)新點(diǎn)如下:(1)提出一種新型的非監(jiān)督神經(jīng)網(wǎng)絡(luò)模型——深度卷積降噪自編碼機(jī),從未標(biāo)記的圖像樣本中學(xué)習(xí)有效的特征表達(dá),從而改善了當(dāng)前主流深度網(wǎng)絡(luò)訓(xùn)練需要大量標(biāo)記樣本的問題。(2)提出了一種新的正則化方法——結(jié)構(gòu)化去相關(guān)約束,能夠有效地規(guī)范化神經(jīng)網(wǎng)絡(luò)模型,防止模型陷入過擬合,同時(shí)使得模型學(xué)習(xí)結(jié)構(gòu)化和不冗余的特征表達(dá),極大地提升了網(wǎng)絡(luò)模型的特征學(xué)習(xí)能力和圖像分類能力。
[Abstract]:In recent years, with the development of imaging technology, image, video and other data are more and more abundant. How to extract effective information from massive data has also become a difficult problem. Deep learning has been proved to be an effective solution. Depth neural network is a multi-level feature learning model. Abstract features can be automatically extracted from the original visual data for subsequent image classification and target detection. However, the training of depth neural networks often requires a large number of labeled samples and a large number of model parameters. In order to solve these problems, an improved depth neural network model and a method to prevent over-fitting are proposed based on the previous models. In order to improve the performance of depth neural network in image recognition, the innovation of this paper is as follows: 1) A new unsupervised neural network model-depth convolution noise reduction self-coding machine is proposed, which can learn effective feature expression in unmarked image samples. Therefore, the problem of large number of tagged samples is improved in the training of current mainstream deep network. A new regularization method, structured decorrelation constraint, is proposed, which can effectively standardize the neural network model and prevent the model from falling into overfitting. At the same time, it makes the model learning structure and non-redundant feature representation, greatly improve the network model feature learning ability and image classification ability.
【學(xué)位授予單位】：武漢大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP391.41;TP18

【相似文獻(xiàn)】

相關(guān)期刊論文 前2條

1 胡宏宇;王殿海;李志慧;曲昭偉;魏巍;;基于視頻的車輛特征表達(dá)與分類算法[J];交通與計(jì)算機(jī);2008年06期

2 ;[J];;年期

相關(guān)博士學(xué)位論文 前1條

1 施建宇;蛋白質(zhì)亞細(xì)胞定位特征表達(dá)與分類算法研究[D];西北工業(yè)大學(xué);2006年

相關(guān)碩士學(xué)位論文 前1條

1 熊維;面向視覺特征表達(dá)的深度學(xué)習(xí)算法研究[D];武漢大學(xué);2017年

，

本文編號：1694376