Unravelling Human Actions with Deep Learning Techniques
Keywords:
Recognition of human actions/activities, deep learning techniques, convolutional neural networks (CNN), MobileNet, Nasnet, Convnext, OpenCV.Abstract
Human activity recognition plays a crucial role in interpersonal communication and relationships as it provides insights into a person's identity, personality, and psychological state. Extracting this information is challenging due to its complex nature. The scientific fields of computer vision and Deep learning extensively study the human ability to recognize activities, leading to the development of various applications such as video surveillance systems, human-computer interaction, and characterization of human behaviour in robotics. Recognizing multiple activities simultaneously is a requirement for many of these applications. Numerous publications have focused on the important field of human activity recognition in images and videos. An OpenCV-based deep learning algorithm, specifically a Convolutional Neural Network (CNN), is proposed in this paper. This calculation can successfully prepare datasets and precisely perceive human activities and exercises.
References
- Activity Recognition with Smartphone Sensors, Xing Su, Hanghang Tong, and Ping Ji, tsinghua science and technology ISSN ll1007-0214ll02/11pp235-249, June 2014, Volume 19.
- Nicholas D.Lane , Emiliano Miluzzo, Hong Lu, Daniel Peebles,‖A Review of Cell Phone Sensing‖, IEEE Interchanges Magazine September 2010
- Activity recognition using the knearest neighbor algorithm on a smartphone with a tri-axial accelerometer, SahakKaghyan and HakobSarukhanyan, International Journal "Information Models and Analyses" Vol. 1 / 2012J.
- Mustafa Kose, OzlemDurmazIncel,CemErsoy ‖Online Human Movement Acknowledgment on Shrewd Phones‖2nd Global Studio on Versatile Detecting, April 16, 2012.
- "Recognizing Human Activities User Independently on Smartphones Based on Accelerometer Data," by PekkaSirrtola and JuhaRöning, is published in the International Journal of Artificial Intelligence and Interactive Multimedia, Vol. 1, No 5.
- Activity Recognition Using Cell Phone Accelerometers, sensor KDD '10, July 25, 2010, Washington, DC, USA, Jennifer R. Kwapisz, Gary M. Weiss, and Samuel A. Moore. Right of use: 2010ACM.
- Activity Recognition from User-Annotated Acceleration Data, by Ling Bao and Stephen S. Intille. Ferscha and F. Mattern (Eds) Inescapable 2004, LNCS 3001, pp. 1–17, 2004. 2004. Springer Berlin Heidelberg
- "Mobile Sensor Data Collector using Android smartphone," by Won-Jae Yi, WeidiJia, and JafarSaniie, Department of Electrical and Computer Engineering, Illinois Institute of Technology, 3301 S. Dearborn St.103SH, Chicago, IL, USA.
- B. Brown, H. Muller, I. Anderson, J. Maitland, S. Sherwood, L. Barkhuus, M. Chalmers, M. Hall, and B. Brown Shakra: following and sharing everyday action levels with unaugmented cell phones. In 2007, Mobile Networks and Applications, 12(2-3), pp. 185-199
- Android. A look at the sensors. 2014. [http://developer.android.com/guide/topics/sensors/sensors_overview.html] Online; [accessed March 1, 2014]
- D. Anguita, A. Ghio, L. Oneto, X. Parra, and J. L. Reyes-Ortiz. A support vector machine with multiple classes that is friendly to hardware is used to identify human activity on smartphones. In Encompassing Helped Residing and Home Consideration, pages 216-223. Springer, 2012.
- Apple. 2014 Nike + iPod Application [ Online; gotten to 01-Walk 2014]
- Apple. Reference to the UIAcceleration Class. https://developer.apple.com/library/ios/documentation/uikit/reference/UIAcceleration_Class/Reference/UIAcceleration.h tml, 2014. [ Online; gotten to 17-Walk 2014]
- W.- Y. Deng, Q.- H. Zheng, and Z.- M. Wang. Recognizing cross-person activities with an extreme learning machine with a reduced kernel Neural System
- Chen Z, Zhu Q, Soh YC, and Zhang L. (2017). CT-PCA and online SVM-based robust human activity recognition with smartphone sensors. 13(6):3070–3080, IEEE Trans IndustrInf
- Cho H, Yoon SM (2018) Separation and vanquish based 1D CNN human action acknowledgment utilizing test information honing. 18 (4): 1055, Sensors
- Cho K, Van Merriënboer B, Gulcehre C, Bahdanau D, Bougares F, Schwenk H, Bengio Y (2014) Learning phrase portrayals utilizing RNN encoder-decoder for measurable machine interpretation. In: Doha, Qatar, Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 1724–1734
- Chung J, Gulcehre C, Cho K, Bengio Y (2014) Experimental assessment of gated intermittent brain networks on succession demonstrating. In: Deep learning workshop at NIPS 2014
- Arcface: Deng J, Guo J, Xue N, and Zafeiriou S (2019) for deep face recognition, additive angular margin loss In: Procedures of the IEEE Meeting on PC Vision and Example Acknowledgment, pp 4690-4699
- Dobbin KK, Simon RM (2011) Ideally dividing cases for preparing and testing high layered classifiers. BMC Drug Genom 4(1):31
- Personalization of classification models for the recognition of human activity by Ferrari A, Micucci D, Mobilio M, and Napoletano P (2020). IEEE Access 8, pages 32066-32079
- Accelerometry-based detection of posture and motion, by Foerster F, Smeja M, and Fahrenberg J a monitoring ambulatory validation study. Comput Murmur Behav 15(5):571-583
- Gao W, Zhang L, Teng Q, Wu H, Min F, He J (2020) DanHAR: double consideration network for multimodal human action acknowledgment utilizing wearable sensors.
- Johnson RA, Miller I, and Freund JE (2000) provided engineers with probability and statistics. Pearson Training, London, p 642
- Jordao A, Kloss R, Schwartz WR (2018) Inert HyperNet: investigating the convolutional neural network's layers. In: IEEE, Proceedings of the 2018 International Joint Conference on Neural Networks (IJCNN), 1–7
Downloads
Published
Issue
Section
License
Copyright (c) IJSRCSEIT

This work is licensed under a Creative Commons Attribution 4.0 International License.