基于層次分類的手機(jī)位置無關(guān)的動作識別
doi: 10.11999/JEIT160253 cstr: 32379.14.JEIT160253
基金項(xiàng)目:
天津市重大科技專項(xiàng)(13ZCZDGX01098)，天津市自然科學(xué)基金(16JCQNJC00700)
Hierarchical Classification-based Smartphone Displacement Free Activity Recognition
Funds:
The Key Project in Tianjin Science Technology Pillar Program (13ZCZDGX01098), The Natural Science Foundation of Tianjin (16JCQNJC00700)
-
摘要: 使用智能手機(jī)中集成的加速度傳感器識別用戶日常動作在慣性定位、個性化推薦、運(yùn)動量評估等領(lǐng)域有重要的應(yīng)用。手機(jī)位置不固定導(dǎo)致的動作識別率低下是該領(lǐng)域面臨的主要問題。為了提高手機(jī)位置不固定時的動作識別率，該文提出一種基于層次分類的動作識別方法。該方法將動作識別分為多層，每一層包含一個分類器。在訓(xùn)練某一層分類器時，首先根據(jù)本層訓(xùn)練樣本集進(jìn)行特征選擇并訓(xùn)練分類器。然后使用訓(xùn)練得到的分類器對訓(xùn)練樣本分類，并計算分類結(jié)果的可信度。最后通過對低可信度的樣本進(jìn)行剪枝得到下層分類器的訓(xùn)練樣本。對未知類別的樣本分類時，首先使用第1層分類器分類。如果分類結(jié)果可信度較高，則分類結(jié)束；否則使用下層分類器分類，直至所有分類器遍歷完。實(shí)驗(yàn)部分通過對采集的動作數(shù)據(jù)進(jìn)行仿真，驗(yàn)證了該文方法的有效性。結(jié)果表明，與單層分類器相比，該方法可以將動作識別率由85.2%提高至89.2%。Abstract: Human activity recognition based on accelerometer embedded in smartphones is wildly applied to inertial positioning, personalized recommendation, daily exercise estimating and other fields. The low recognition rate which caused by varying phone displacement is a crucial problem which needs to solve. To improve the recognition rate when the phones displacement is unfixed, a hierarchical classification-based activity recognition method is proposed. The activity recognition process is divided into multiple layers in this method, and each layer contains a classifier. For training each layers classifier, it runs the feature selection algorithm first, and the classifier is trained based on the selected features. Then, the trained classifier is used to classify the training set, and each samples classification confidence is calculated. Finally, samples whose confidence is lower than the hierarchical threshold are selected as the next layers training set. This process continues until each activitys sample number is less than the predefined pruning threshold. When an unlabeled sample comes, the first layer is used to classify this sample. If the classification confidence is higher than the hierarchical threshold, the recognition is over. Otherwise, the next layer will repeat this process until all the layers are traversed. The experiment collects activity data, and simulates the activity recognition. The simulation show that compared with the current methods, this method may improve the recognition rate from 85.2% to 89.2%.
-
Key words:
- Activity recognition /
- Accelerometer /
- Hierarchical classification /
- Feature selection
-
ANDREAS Bulling, ULF Blanke, and BERNT Schiele. A tutorial on human activity recognition using body-worn inertial sensors[J]. ACM Computing Surveys, 2014, 46(3): 1-33. CHIANG Junghsien, YANG Peiching, and TU Hsuan. Pattern analysis in daily physical activity data for personal health management[J]. Pervasive and Mobile Computing, 2014, 13: 13-25. SINZIANA Mazilu, ULF Blanke, MORAN Dorfman, et al. A wearable assistant for gait training for parkinsons disease with freezing of gait in out-of-the-lab Environments[J]. ACM Transactions on Interactive Intelligent Systems, 2015, 5(1): 1-5. PARK Jungeun, AIM Patel, DOROTHY Curtis, et al. Online pose classification and walking speed estimation using handheld devices[C]. The 2012 ACM Conference on Ubiquitous Computing, Pittsburgh, PA, USA, 2012: 113-122. JORGE Luis Reyes Ortiz. Smartphone-based human activity recognition[D]. [Ph.D. dissertation], Polytechnic University of Catalonia, 2014: 97-114. CHEN Yiqiang, ZHAO Zhongtang, WANG Shuangquan, et al. Extreme learning machine-based device displacement free activity recognition model[J]. Soft Computing, 2012, 16(9): 1617-1625. WANG Changhai, ZHANG Jianzhong, LI Meng, et al. A smartphone location independent activity recognition method based on the angle feature[C]. International Conference on Algorithms and Architectures for Parallel Processing, Dalian, China, 2014: 179-191. SEYED Amir Hoseini-tabatabaei, ALEXANDER Gluhak, and RAHIM Tafazolli. A survey on smartphone-based systems for opportunistic user context recognition[J]. ACM Computing Surveys, 2013, 45(3): 1-27. 張靜, 宋銳, 郁文賢, 等. 基于混淆矩陣和 Fisher 準(zhǔn)則構(gòu)造層次化分類器[J]. 軟件學(xué)報, 2005, 16(9): 1560-1567. ZHANG Jing, SONG Rui, YU Wenxian, et al. Construction of hierarchical classifiers based on the confusion matrix and Fishers principle[J]. Journal of Software, 2005, 16(9): 1560-1567. 高洪波, 王洪玉, 劉曉凱. 一種基于分層學(xué)習(xí)的關(guān)鍵點(diǎn)匹配算法[J]. 電子與信息學(xué)報, 2013, 35(11): 2751-2757. doi: 10.3724/SP.J.1146.2013.00347. GAO Hongbo, WANG Hongyu, and LIU Xiaokai. A keypoint matching method based on hierarchical learning[J]. Journal of Electronics Information Technology, 2013, 35(11): 2751-2757. doi: 10.3724/SP.J.1146.2013.00347. 張翔, 鄧趙紅, 王士同, 等. 極大熵 Relief 特征加權(quán)[J]. 計算機(jī)研究與發(fā)展, 2015, 48(6): 1038-1048. ZHANG Xiang, DENG Zhaohong, WANG Shitong, et al. Maximum entropy relief feature weighting[J]. Journal of Computer Research and Development, 2015, 48(6): 1038-1048. XUE Yang and JIN Lianwen. A naturalistic 3D acceleration- based activity dataset benchmark evaluations[C]. International Conference on Systems Man and Cybernetics, Istanbul, Turkey, 2010: 4081-4085. ATTILA Reiss. Personalized mobile physical activity monitoring for everyday life[D]. [Ph.D. dissertation], Technical University of Kaiserslautern, 2014: 111-127. HENAR Martin, ANA M Bermardos, JOSUE Iglesias, et al. Activity logging using lightweight classification techniques in mobile devices[J]. Personal and Ubiquitous Computing, 2013, 17(4): 675-695. -
計量
- 文章訪問數(shù): 1210
- HTML全文瀏覽量: 158
- PDF下載量: 378
- 被引次數(shù): 0
下載:
