本文選題：人體行為識別 + 可穿戴式動作捕捉系統(tǒng)　； 參考：《安慶師范大學(xué)》2017年碩士論文

【摘要】：人體行為識別已經(jīng)成為人機(jī)交互和普適計算的關(guān)鍵研究領(lǐng)域,其目標(biāo)是提供關(guān)于用戶行為的信息,允許計算機(jī)主動地幫助用戶完成任務(wù)。計算機(jī)視覺研究一直處于這一工作的前沿。隨著傳感器技術(shù)的發(fā)展,研究人員逐漸開始使用慣性傳感器(如加速度計或陀螺儀)進(jìn)行行為識別研究。因此,基于無線傳感器的人體行為識別方法正成為研究熱點。特征提取是人體行為識別的核心問題之一,提取最有效的特征是提高人體行為識別精度的重要途徑�，F(xiàn)有的基于傳感器的人體行為識別方法,在特征提取階段往往采用諸如加速度或角速度數(shù)據(jù)的均值,方差,峰度等離散數(shù)據(jù)特征。一方面,這些離散型特征并不能反映人體運(yùn)動的連續(xù)性,另一方面,這些通過手工選擇的特征依賴于先驗知識。因此,本文主要針對以上不足,采用基于傳感器的可穿戴式動作捕捉系統(tǒng),研究適用于人體行為識別的運(yùn)動特征。其一,人體運(yùn)動過程中,肢體的運(yùn)動是連續(xù)的,而對應(yīng)的動作捕捉數(shù)據(jù)是離散的。為了更好地分析人體日常行為的連續(xù)性與周期性,本文在第二章中提出了一種基于函數(shù)型數(shù)據(jù)分析的人體行為識別方法。首先,利用函數(shù)型數(shù)據(jù)分析方法,將可穿戴式動作捕捉系統(tǒng)采集的人體周期行為數(shù)據(jù)函數(shù)化,通過函數(shù)準(zhǔn)確地定義數(shù)據(jù)的連續(xù)性與周期性。然后,從動作捕捉數(shù)據(jù)的函數(shù)中分別提取周期數(shù)據(jù)特征與函數(shù)型特征,并結(jié)合支持向量機(jī)方法,有效地識別了多類別的人體日常行為。其二,深度神經(jīng)網(wǎng)絡(luò)能夠自動地學(xué)習(xí)低層次到高層次的分布式特征,并以此自學(xué)習(xí)特征代替?zhèn)鹘y(tǒng)方法中手工提取的特征。本文在第三章中定義了短時行為的概念,并基于短時行為提出了一種基于深度學(xué)習(xí)的人體行為識別方法。首先,利用滑動窗口分割的方法,構(gòu)建了一個包含不同人體短時行為模式的過完備模式庫。然后,基于此過完備模式庫,使用卷積神經(jīng)網(wǎng)絡(luò)實現(xiàn)短時行為特征的自動學(xué)習(xí),并最終實現(xiàn)對短時日常行為的分類識別。
[Abstract]:Human behavior recognition has become a key research field in human-computer interaction and pervasive computing. Its goal is to provide information about user behavior and to allow computers to actively help users complete tasks. Computer vision research has been at the forefront of this work. With the development of sensor technology, researchers began to use inertial sensors (such as accelerometers or gyroscopes) to identify behavior. Therefore, human behavior recognition based on wireless sensor is becoming a hot research topic. Feature extraction is one of the core problems in human behavior recognition. The most effective feature extraction is an important way to improve the accuracy of human behavior recognition. The existing sensor-based human behavior recognition methods often use discrete data features such as the mean variance kurtosis of acceleration or angular velocity data in feature extraction stage. On the one hand, these discrete features can not reflect the continuity of human motion, on the other hand, these features selected by hand depend on prior knowledge. Therefore, in this paper, a wearable motion capture system based on sensors is used to study the motion features of human behavior recognition. First, the body motion is continuous, and the corresponding motion capture data is discrete. In order to better analyze the continuity and periodicity of human daily behavior, a method of human behavior recognition based on functional data analysis is proposed in the second chapter. Firstly, the data of human periodic behavior collected by wearable motion capture system are functioned by the method of functional data analysis, and the continuity and periodicity of the data are accurately defined by the function. Then, the periodic data features and the functional features are extracted from the function of motion capture data, and the support vector machine (SVM) method is used to identify the human daily behavior effectively. Secondly, the deep neural network can automatically learn the distributed features from the lower level to the high level, and use the self-learning features to replace the features extracted by hand in the traditional methods. In the third chapter, the concept of short-term behavior is defined, and a human behavior recognition method based on deep learning is proposed based on short-term behavior. Firstly, an overcomplete pattern library containing different human short term behavior patterns is constructed by using sliding window segmentation method. Then, based on the over-complete pattern library, the convolution neural network is used to realize the automatic learning of short-time behavior features, and finally to realize the classification and recognition of short-time daily behavior.
【學(xué)位授予單位】：安慶師范大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP212.9;TP391.4

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 王啟寧;鄭恩昊;陳保君;麥金耿;;面向人機(jī)融合的智能動力下肢假肢研究現(xiàn)狀與挑戰(zhàn)[J];自動化學(xué)報;2016年12期

2 常亮;鄧小明;周明全;武仲科;袁野;楊碩;王宏安;;圖像理解中的卷積神經(jīng)網(wǎng)絡(luò)[J];自動化學(xué)報;2016年09期

3 朱煜;趙江坤;王逸寧;鄭兵兵;;基于深度學(xué)習(xí)的人體行為識別算法綜述[J];自動化學(xué)報;2016年06期

4 李鋒;潘敬奎;;基于三軸加速度傳感器的人體運(yùn)動識別[J];計算機(jī)研究與發(fā)展;2016年03期

5 蘇本躍;陳曉慧;童星慧;王廣軍;;兩類函數(shù)型數(shù)據(jù)主成分分析方法及其應(yīng)用[J];統(tǒng)計與決策;2015年17期

6 王德青;劉曉葳;朱建平;;基于自適應(yīng)迭代更新的函數(shù)型數(shù)據(jù)聚類方法研究[J];統(tǒng)計研究;2015年04期

7 佟麗娜;侯增廣;彭亮;王衛(wèi)群;陳翼雄;譚民;;基于多路sEMG時序分析的人體運(yùn)動模式識別方法[J];自動化學(xué)報;2014年05期

8 姜高霞;王文劍;;經(jīng)濟(jì)周期波動的函數(shù)型時序分解方法——基于CPI的實證分析[J];統(tǒng)計與信息論壇;2014年03期

9 蘇為華;孫利榮;崔峰;;一種基于函數(shù)型數(shù)據(jù)的綜合評價方法研究[J];統(tǒng)計研究;2013年02期

10 肖玲;李仁發(fā);羅娟;;體域網(wǎng)中一種基于壓縮感知的人體動作識別方法[J];電子與信息學(xué)報;2013年01期

，

本文編號：1950598