Abstract
Human motion recognition has attracted the attention of many global media platforms in video feature analysis, and has evolved into a branch of AI technology. In order to understand the human motion recognition process, improve the recognition efficiency of human motion, and provide reference for future researchers. This paper studies the human motion recognition system based on LSTM neural system, mainly referring to the action recognition theory of Facebook, and expanding the types and scope of the recognized images. This system uses the bottom-up idea to recognize the human body in the picture or image set. Firstly, human joints are identified, and then multiple joints are combined as a node in the system. Due to the use of LSTM, this method can recognize the actions of different regions without human recognition, and then combine them. In theory, this method does not rely on the recognition of human body, so the recognition speed and accuracy will be improved. The research shows that the algorithm based on LSTM recurrent neural network achieves a high recognition rate. In the experiments of human motion information database, four groups of recognition rates are higher than the three methods, and the average recognition rate is 92.64%, higher than the other three methods. In the three groups of cross validation experiments, the average recognition rate is 86.97%, higher than the four methods. The recognition rate of the utkinect action 3D database is 95.96%, indicating that the deep learning algorithm has the ability to extract deep features of human action features that are useful for action recognition.
Similar content being viewed by others
Data availability statements
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
References
Abbas F, Feng D, Habib S et al (2018) Short term residential load forecasting: an Improved optimal nonlinear auto regressive (NARX) method with exponential weight decay function. Electronics 7(12):432
Cao Y, Wang F, Lu X et al (2020) Contactless body movement recognition during sleep via WiF i signals. IEEE Internet Things J 7(3):2028–2037
Dai J, Song H, Sheng G et al (2018) Prediction method for power transformer running state based on LSTM network. Gaodianya Jishu/high Voltage Eng 44(4):1099–1106
Ghasemzadeh H, Jafari R (2019) Physical movement monitoring using body sensor networks: a phonological approach to construct spatial decision trees. IEEE Trans Industr Inf 7(1):66–77
Goldberg Y (2017) Neural network methods for natural language processing. Synth Lect Hum Lang Technol 10(1):301–309
Gurung S, Ghose MK, Subedi A (2019) Deep learning approach on network intrusion detection system using NSL-KDD dataset. Int J Comput Netw Inf Secur 11(3):8–14
Kesiman MWA, Dermawan KT (2021) AKSALont: Automatic transliteration application for Balinese palm leaf manuscripts with LSTM Model. Jurnal Teknologi Dan Sistem Komputer 9(3):142–149
Kim J, Shin Y, Choi E (2019) An intrusion detection model based on a convolutional neural network. J Multimed Inf Syst 6(4):165–172
Lin F, Lin Y, Lv X et al (2018) Short-term parallel power load forecasting based on balanced KNN. Electric Power 51(10):88–94 (102)
Liu J, Shahroudy A, Xu D et al (2018a) Skeleton-based action recognition using spatio-temporal LSTM network with trust gates. IEEE Trans Pattern Anal Mach Intell 40(12):3007–3021
Liu J, Wang G, Duan LY et al (2018b) Skeleton-based human action recognition with global context-aware attention LSTM networks. IEEE Trans Image Process 27(99):1586–1599
Lloyd JR (2019) GEFCom2012 hierarchical load forecasting: gradient boosting machines and Gaussian processes. Int J Forecast 30(2):369–374
Shen XT, Yu S, Dong Y et al (2020) Human motion recognition method based on deep learning. Comput Eng Des 41(4):1153–1157
Tang C, Wang WJ, Zhang C et al (2019) Human behavior recognition based on RGB-D image features. Pattern Recogn Artif Intell 32(10):901–908
Tian LF, Wu QC, Du QL et al (2019) Recognition of passengers’ abnormal behavior on the escalator based on human skeleton sequence. J South China Univ Technol 47(4):10–19
Wang T, Wang Z, Zhang D et al (2016) Recognizing parkinsonian gait pattern by exploiting fine-grained movement function features. ACM Trans Intell Syst Technol 8(1):1–22
Wang Y, Wu K, Ni LM (2017) WiFall: device-free fall detection by wireless networks. IEEE Trans Mob Comput 16(2):581–594
Wang Z, Zhao B, Ji W et al (2019) Short-term load forecasting method based on GRU-NN model. Autom Elect Power Syst 43(5):86–95
Wu J (2019) Unsupervised intrusion detection based on feature selection. Comput Eng Appl 47(26):79–82
Xu Y, Hou ZJ, Liang JZ et al (2018) Human action recognition with multi-feature fusion by depth image and skeleton data. J Chin Comput Syst 39(8):1865–1870
Yang ZY, Li YC, Yang JC et al (2019) Action recognition with spatio–temporal visual attention on skeleton image sequences. IEEE Trans Circuits Syst Video Technol 29(8):2405–2415
Yu GU, Yantong WANG, Zhi LIU et al (2020) SleepGuardian: an RF-based healthcare system guarding your sleep from afar. IEEE Network 34(2):164–171
Zhang SY, Yang Y, Xiao J et al (2018) Fusing geometric features for skeleton-based action recognition using multilayer LSTM networks. IEEE Trans Multimedia 20(9):2330–2343
Zhang Y, Ai Q, Lin L et al (2019) A very short-term load forecasting method based on Deep LSTM RNN at zone level. Power Syst Technol 43(6):29–37
Acknowledgements
This work was supported by 1. Hunan Province Education Science “the 13th Five-Year Plan” 2019 general project project (Project Number: XJK19BTW004), 2. Hunan Social Science achievements Review Committee 2020 project (Project Number: XSP20YBC119), 3. Ministry of Education Humanities and Social Sciences Foundation Project, (Project Number: 71802119), 4. Natural Science Foundation Fund Youth Project (Project Number: 18YJC620002), 5. Subject of Education Department of Hunan Province (Project Number: 19C0740)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
li, X., Cao, X. Human motion recognition information processing system based on LSTM Recurrent Neural Network Algorithm. J Ambient Intell Human Comput 14, 8509–8521 (2023). https://doi.org/10.1007/s12652-021-03614-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12652-021-03614-x