Deep learning for human activity recognition: A resource efficient implementation on low-power devices
The DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Abstract
Human Activity Recognition provides valuable contextual information for wellbeing, healthcare, and sport applications. Over the past decades, many machine learning approaches have been proposed to identify activities from inertial sensor data for specific applications. Most methods, however, are designed for offline processing rather than processing on the sensor node. In this paper, a human activity recognition technique based on a deep learning methodology is designed to enable accurate and real-time classification for low-power wearable devices. To obtain invariance against changes in sensor orientation, sensor placement, and in sensor acquisition rates, we design a feature generation process that is applied to the spectral domain of the inertial data. Specifically, the proposed method uses sums of temporal convolutions of the transformed input. Accuracy of the proposed approach is evaluated against the current state-of-the-art methods using both laboratory and real world activity datasets. A systematic analysis of the feature generation parameters and a comparison of activity recognition computation times on mobile devices and sensor nodes are also presented.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The proposed method results in an overall accuracy of 92.6%, outperforming other methods such as HMMs [23], DBNs [6], and CNNs [11], as shown in Figure 12b.”
From Deep Recurrent Neural Networks for Human Activity Recognition (2017) · cites this paper as [11]
In the citing paper
“Ravi et al. [170] used a shallow three-layer CNN network including a convolutional layer, a fully connected layer, and a softmax layer to perform on-device activity recognition on a resource-limited platform and shown its effectiveness and efficiency on public datasets.”
From Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances (2022) · cites this paper as [170]
In the citing paper
“The main disadvantage of AIDS is the false alarm rate (FAR), making it difficult to distinguish between normal and abnormal intrusion detection patterns [197].”
From Comparative Analysis of Intrusion Detection Systems and Machine Learning Based Model Analysis through Decision Tree (2023) · cites this paper as [197]
Citations
Cites
The Design and Implementation of FFTW3
Frigo, M., & Johnson, S. G. (2005) · Proceedings of the IEEE · doi:10.1109/jproc.2004.840301 · Cited by 5,086 · cited 1×
“For both architectures, we use the FFTW3 library [17] to extract the spectrogram and the Torch framework [18] to implement the deep learning model.”
MethodActivity recognition using cell phone accelerometers
Kwapisz, J. R., Weiss, G. M., & Moore, S. A. (2011) · ACM SIGKDD Explorations Newsletter · doi:10.1145/1964897.1964918 · Cited by 2,912 · cited 1×
“1 6 29 Front trouser pocket 20Hz 1,098,207 [19] (thigh) Daphnet FoG 2 10 Trunk, thigh, ankle 64Hz 1,917,887 [20] Skoda 10 1 Arms (20 positions) 98Hz ∼ 701, 440 [21]”
A Survey on Human Activity Recognition using Wearable Sensors
Lara, O. D., & Labrador, M. A. (2013) · IEEE Communications Surveys & Tutorials · doi:10.1109/surv.2012.110112.00192 · Cited by 2,662 · cited 1×
“1) Raw inertial data is highly fluctuating over time, making classification impossible when using a single data point in time [15].”
BackgroundA symbolic representation of time series, with implications for streaming algorithms
Lin, J., Keogh, E., Lonardi, S., & Chiu, B. (2003) · Proceedings of the 8th ACM SIGMOD workshop on Research issues in data mining and knowledge discovery · doi:10.1145/882082.882086 · Cited by 1,799 · cited 3×
“Features, such as mean [1], Fourier transforms [2], and symbols [3], are typically extracted from segments of data and then trained using classification methods [4]–[8].”
BackgroundMethodA tutorial on human activity recognition using body-worn inertial sensors
Bulling, A., Blanke, U., & Schiele, B. (2014) · ACM Computing Surveys · doi:10.1145/2499621 · Cited by 1,668 · cited 3×
“Features, such as mean [1], Fourier transforms [2], and symbols [3], are typically extracted from segments of data and then trained using classification methods [4]–[8].”
BackgroundMethod(2015) · Cited by 989
Convolutional Neural Networks for Human Activity Recognition using Mobile Sensors
Zeng, M., Nguyen, L. T., Yu, B., Mengshoel, O. J., Zhu, J., Wu, P., & Zhang, J. (2014) · Proceedings of the 6th International Conference on Mobile Computing, Applications and Services · doi:10.4108/icst.mobicase.2014.257786 · Cited by 876 · cited 2×
“In deep learning, features are abstracted automatically from the data instead of being handcrafted, which allows these machine learning methods to be more effectively used across a range of different classification tasks [10]– [13].”
MethodTime–frequency feature representation using energy concentration: An overview of recent advances
Sejdić, E., Djurović, I., & Jiang, J. (2009) · Digital Signal Processing · doi:10.1016/j.dsp.2007.12.004 · Cited by 756 · cited 1×
“STFT is used to determine the sinusoidal frequency and phase content of local sections of a signal as it changes over time [16].”
MethodWearable Assistant for Parkinson’s Disease Patients With the Freezing of Gait Symptom
Bachlin, M., Plotnik, M., Roggen, D., Maidan, I., Hausdorff, J. M., Giladi, N., & Troster, G. (2010) · IEEE Transactions on Information Technology in Biomedicine · doi:10.1109/titb.2009.2036165 · Cited by 693 · cited 2×
“We obtain comparable accuracy results even against approaches that are more resource demanding [7], [13], [20], [21].”
Result(2006) · doi:10.1007/1-84628-484-8 · Cited by 667
Self-citation
Research works printed but not linked: 1
- walse, 2016, A study of human activity recognition using AdaBoost classifiers on WISDM dataset, The Institute of Integrative Omics and Applied Biotechnology Journal, 7, 68
Cited by
Items 101–200 of 311
Detection of human activities using multi-layer convolutional neural network
Abdellatef, E., Al‐Makhlasawy, R. M., & Shalaby, W. A. (2025) · Scientific Reports · doi:10.1038/s41598-025-90307-6 · Cited by 34
“Ravi et al. 20 present a deep learning-based human activity detection technique to provide an accurate and real-time categorization for wearable low-power devices.”
Balanced Neural Architecture Search and Its Application in Specific Emitter Identification
Du, M.-Y., He, X., Cai, X., & Bi, D.-P. (2021) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2021.3107633 · Cited by 28
“Models like convolutional neural network (CNN) and recurrent neural network (RNN) employ a data-driven approach to learn discriminative features from raw sensor data to infer complex, sequential and contextual information in a hierarchical manner [17].”
Vuong, T. H., Doan, T., & Takasu, A. (2023) · Sensors · doi:10.3390/s23249721 · Cited by 33
“DNN [17,38,39] are then used to classify each spectrum as an activity.”
MethodA Framework for the Capture and Analysis of Product Usage Data for Continuous Product Improvement
Voet, H. B., Altenhof, M., Ellerich, M., Schmitt, R. H., & Linke, B. S. (2018) · Journal of Manufacturing Science and Engineering · doi:10.1115/1.4041948 · Cited by 33
A Smartphone Lightweight Method for Human Activity Recognition Based on Information Theory
Bragança, H. L. D. S., Colonna, J. G., Lima, W. S., & Souto, E. (2020) · Sensors · doi:10.3390/s20071856 · Cited by 32
“These features are simple to understand and have low computational complexity, very suitable for mobile devices [7,13].”
BackgroundSmartphone Sensor-Based Human Activity Recognition Robust to Different Sampling Rates
Hasegawa, T. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.3038281 · Cited by 32
“’s study [25] collected data in an SR mixed”
BackgroundInfluentialZhang, H., Chen, L., Zhang, Y., Hu, R., He, C., Tan, Y., & Zhang, J. (2022) · Journal of Sensors · doi:10.1155/2022/8507706 · Cited by 32
Intelligent system for human activity recognition in IoT environment
Khaled, H., Abu-Elnasr, O., Elmougy, S., & Tolba, A. S. (2021) · Complex & Intelligent Systems · doi:10.1007/s40747-021-00508-5 · Cited by 31
“In the second place, Spatio-Temporal Deep Learning [46] has accuracy of 98.53%, in third-place Deep learning low power device [41] has accuracy of 98.2% while in the third-place, CNN + BLSTM [44] has accuracy of 97.8%.”
BackgroundInfluentialDeep learning of smartphone sensor data for personal health assistance
Li, H., & Trocan, M. (2019) · Microelectronics Journal · doi:10.1016/j.mejo.2018.01.015 · Cited by 30
Machine Learning for Detection and Risk Assessment of Lifting Action
Thomas, B., Lu, M., Jha, R., & Bertrand, J. (2022) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2022.3212666 · Cited by 30
“Ravì et al. [16] introduced a low-power model for real-time classification on wearable devices.”
BackgroundWan, J., Xu, S., Lin, J., Ji, L., Cheng, J., Li, Z., Qu, X., & Li, C. (2025) · Small · doi:10.1002/smll.202504078 · Cited by 29
Deep generative cross-modal on-body accelerometer data synthesis from videos
Zhang, S., & Alshurafa, N. (2020) · Adjunct Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers · doi:10.1145/3410530.3414329 · Cited by 21
“ple, for physical activity recognition, most commonly used public datasets only contain labeled sensor data ranging from 15 hours (in 29 subjects) [12] to 30 hours (in 10 subjects) [19], and it has been shown that increasing the amount of data used to train the 223”
Resource-Efficient Continual Learning for Sensor-Based Human Activity Recognition
Leite, C. F. S., & Xiao, Y. (2022) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3530910 · Cited by 29
Self-Gated Recurrent Neural Networks for Human Activity Recognition on Wearable Devices
Vu, T. H., Dang, A., Dung, L., & Wang, J. (2017) · Thematic Workshops of ACM Multimedia · doi:10.1145/3126686.3126764 · Cited by 28
“In [9, 10, 14, 19, 31], convolutional neural networks (CNN) are successfully applied to the problem.”
MethodLamooki, S. R., Hajifar, S., Kang, J., Sun, H., Megahed, F. M., & Cavuoto, L. A. (2022) · Applied Ergonomics · doi:10.1016/j.apergo.2022.103732 · Cited by 27
Deep Learning-Based Video Action Recognition: A Review
Wani, M. H., & Faridi, A. R. (2022) · International Conference on Computing, Communication, and Intelligent Systems (ICCCIS) · doi:10.1109/icccis56430.2022.10037736 · Cited by 27
Human Activity Recognition based on Real Life Scenarios
Machot, F. A., Ranasinghe, S., Plattner, J., & Jnoub, N. (2018) · IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops) · doi:10.1109/percomw.2018.8480138 · Cited by 27
“Other works like in [10] proposed a technique based on a deep learning methodology to offer real-time activity recognition to be applied to limited power devices.”
Jeong, C. Y., & Kim, M. (2019) · Sensors · doi:10.3390/s19173688 · Cited by 26
“[15] also proposed a neural network that combined handcrafted features and the features extracted from a shallow network [14].”
BackgroundMethodHuman Activity Recognition Through Ensemble Learning of Multiple Convolutional Neural Networks
Zehra, N., Azeem, S. H., & Farhan, M. (2021) · Conference on Information Sciences and Systems · doi:10.1109/ciss50987.2021.9400290 · Cited by 26
“[11] introduces a method that takes into account the increased computational complexities and computing resource overheads tied with many deep learning approaches.”
MethodUnderstanding and Improving Deep Neural Network for Activity Recognition
Xue, L., Si, X., Nie, L., Jiazhen, L., Renjie, D., Zhan, D., & Dianhui, C. (2018) · EAI International Conference on Mobile Multimedia Communications (MobiMedia) · doi:10.4108/eai.21-6-2018.2276632 · Cited by 25
Human Activity Recognition Based on Acceleration Data From Smartphones Using HMMs
Iloga, S., Bordat, A., Kernec, J. L., & Romain, O. (2021) · IEEE Access · doi:10.1109/access.2021.3117336 · Cited by 26
“the use of publicly available datasets to enable further comparisons [17]–[25], [28], [33], [35], [36], [43], [49], [51], [52], [54]–[59], [61], [62], [64], [67], [72], [74]–[81], [83], [86]–[88].”
IMU-Based Movement Trajectory Heatmaps for Human Activity Recognition
Konak, O., Wegner, P., & Bert Arnrich, A. (2020) · Sensors · doi:10.3390/s20247179 · Cited by 24
“[17] applied a short-time discrete Fourier transform (STDFT) to time-series signals for constructing a time-frequency-spectral image.”
MethodArtificial neural networks for human activity recognition using sensor based dataset
Geravesh, S., & Rupapara, V. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13716-z · Cited by 24
“In research work, [20], a HAR technique based is designed to enable real-time and accurate classification for low-power wearable sensors based on a deep learning methodology.”
MethodKundu, S., Mallik, M., Saha, J., & Chowdhury, C. (2024) · International Journal of Information Technology · doi:10.1007/s41870-024-02305-y · Cited by 24
“In the deep learning model proposed in [39], feature extraction was done by using a temporal convolution on the spectrogram domain (representing with frequency and time) of the gyroscope and accelerometer data, considering fixed window size.”
Online Learning of Wearable Sensing for Human Activity Recognition
Zhang, Y., Gao, B., Yang, D., Woo, W. L., & Wen, H. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3188785 · Cited by 23
“The use of complicated machine-learning-based methods such as DNNs need to transfer the data to the server with the help of bluetooth or WiFi [34], [35].”
MethodHARNet: Towards On-Device Incremental Learning using Deep Ensembles on Constrained Devices
Sundaramoorthy, P., Gudur, G. K., Moorthy, M. R., Bhandari, R. N., & Vijayaraghavan, V. (2018) · International Workshop on Embedded and Mobile Deep Learning (EMDL) · doi:10.1145/3212725.3212728 · Cited by 23
“[14]performedtemporalconvolutionsonShort-Time Fourier Transform (STFT) spectrogram of the input signals.”
MethodConformer-Based Human Activity Recognition Using Inertial Measurement Units
Seenath, S., & Dharmaraj, M. (2023) · Sensors · doi:10.3390/s23177357 · Cited by 22
“Data Availability Statement: Sensor-based HAR datasets including WISDM [34] and USCHAD [35], which are openly accessible for use in research, were employed in the studies.”
MethodAli, A. M., & Abdelhafeez, A. (2022) · Sustainable Machine Intelligence Journal · doi:10.61185/smij.2022.8463 · Cited by 23
Andreu-Pérez, J., Garcia‐Gancedo, L., McKinnell, J., van der Drift, A., Powell, A., Hamy, V., Keller, T., & Yang, G. (2017) · Sensors · doi:10.3390/s17092113 · Cited by 22
“is one-dimensional, some works have proposed using a time frequency representation as input [42].”
BackgroundSelf-citationBlock-Wise Training Residual Networks on Multi-Channel Time Series for Human Activity Recognition
Teng, Q., Zhang, L., Tang, Y., Song, S., Wang, X., & He, J. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3085360 · Cited by 21
Jaberi, M., & Ravanmehr, R. (2022) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2022.101620 · Cited by 21
Energy expenditure estimation using visual and inertial sensors
Tao, L., Burghardt, T., Mirmehdi, M., Damen, D., Cooper, A. R., Camplani, M., Hannuna, S., Paiement, A., & Craddock, I. J. (2017) · IET Computer Vision · doi:10.1049/iet-cvi.2017.0112 · Cited by 21
“Approaches based on deep learning are currently being explored to create more generalised learning methods that generate features directly from the input data and promise to optimise performance further [28].”
BackgroundImpact of Data Preparation in Freezing of Gait Detection Using Feature-Less Recurrent Neural Network
Esfahani, A. H., Dyka, Z., Ortmann, S., & Langendörfer, P. (2021) · IEEE Access · doi:10.1109/access.2021.3117543 · Cited by 20
“with one convolutional layer followed by a fully connected layer [25].”
BackgroundWhy machine learning (ML) has failed physical activity research and how we can improve
Fuller, D., Ferber, R., & Stanley, K. G. (2022) · BMJ Open Sport & Exercise Medicine · doi:10.1136/bmjsem-2021-001259 · Cited by 20
Smartphone-Based Human Activity Recognition Using CNN in Frequency Domain
Jiang, X., Lu, Y., Lu, Z., & Zhou, H. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-01298-4_10 · Cited by 19
“[17] propose a deep learning architecture: a filter is applied to the input, and the weighted sums are computed in the temporal convolution layer which is followed by a fully-connected layer and a soft-max layer for classification.”
MethodTeng, Q., Tang, Y., & Hu, G. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3364187 · Cited by 18
A Localized Learning Approach Applied to Human Activity Recognition
Amer, A. Y. A., Aerts, J., Vanrumste, B., & Luca, S. (2020) · IEEE Intelligent Systems · doi:10.1109/mis.2020.2964738 · Cited by 19
“We will proceed by introducing our hybrid KNN-LS-SVM algorithm.”
MethodDeep learning for internet of things in fog computing: Survey and Open Issues
Tmamna, J., Ayed, E. B., & Ayed, M. B. (2020) · International Conference on Advanced Technologies for Signal and Image Processing · doi:10.1109/atsip49331.2020.9231685 · Cited by 18
A Review on Human Activity Recognition using Body Sensor Networks
Kumari, K., Chandankhede, P. H., Titarmare, A., Vidhale, B. R., & Tadse, S. (2021) · International Conference on Computing Methodologies and Communication (ICCMC) · doi:10.1109/iccmc51019.2021.9418354 · Cited by 18
“Benny Lo and Guang-Zhong Yang, Daniele Rav`ı, Charence Wong [1], Deep Learning for Human Activity Recognition: A Resource Efficient Implementation on Low-Power Devices: in this paper we have studied that in recent years there is an extensive growth in using deep learning for different classifications because the patte…”
BackgroundYoosefzadeh-Najafabadi, M. (2021) · The Atrium (University of Guelph) · doi:10.13140/rg.2.2.35568.38404 · Cited by 18
Efficient Human Activity Recognition on Wearable Devices Using Knowledge Distillation Techniques
Gonçalves, P. H. N., Bragança, H. L. D. S., & Souto, E. (2024) · Electronics · doi:10.3390/electronics13183612 · Cited by 18
Robust Human Activity Recognition via Wearable Sensors Using Dynamic Gaussian Kernel Learning
Wang, S., Sun, Y., Sha, Y., Yang, G., Cheng, D., Zhang, L., Wu, H., & Song, A. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3355704 · Cited by 17
Self-citationAkıllı ve Geleneksel Giyilebilir Sağlık Cihazlarında Nesnelerin İnterneti
Öcal, H., Doğru, İ. A., & Barışçı, N. (2018) · Journal of Polytechnic · doi:10.2339/politeknik.450290 · Cited by 17
“[76]’ daki çalışmalarından hareketle Rav`ı ve arkadaşları, IoT tabanlı giyilebilir cihazlarda düğüm sensöründen elde edilen insan aktivitelerini belirleyen (Human Activation Recognation-HAR) verileri sınıflama sorununu Derin Öğrenme (Deep Learning) ve Sığ Özellik(Shallow Feature) yöntemini bir arada kullanarak”
InfluentialDeepika, J., & Akilandeswari, J. (2023) · International Conference on Smart Systems and Inventive Technology (ICSSIT) · doi:10.1109/icssit55814.2023.10061005 · Cited by 16
“The accelerations and velocities for fixing threshold and Support Vector Machine (SVM) algorithms utilized as classifier, were used to recognise the falls of individuals in [1].”
MethodInertial-Vision: Cross-Domain Knowledge Transfer for Wearable Sensors
Abebe, G., & Cavallaro, A. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.165 · Cited by 16
“The top of Table 4 (Individual) evaluates the individual classification outputs of feature groups from the ActiveMiles and the WISDM-v2.”
BackgroundMethodResultInfluentialDeep Learning in IoT: Introduction, Applications, and Perspective in the Big Data Era
Devika, G., & Karegowda, A. G. (2020) · Advances in computational intelligence and robotics book series · doi:10.4018/978-1-7998-2108-3.ch001 · Cited by 16
Lakoju, M., Ajienka, N., Khanesar, M. A., Burnap, P., & Branson, D. T. (2021) · Sensors · doi:10.3390/s21154991 · Cited by 16
Bassoli, M., Bianchi, V., & Munari, I. D. (2020) · Sensors · doi:10.3390/s20051362 · Cited by 15
“The development of wearable devices leads to implementation of ML algorithms directly on board [18,19], allowing for the reduction of the amount of data to be transmitted, and with consistent advantages in terms of power consumption and system usability [14].”
MethodError-Tolerant Computation for Voting Classifiers With Multiple Classes
Liu, S., Reviriego, P., Montuschi, P., & Lombardi, F. (2020) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2020.3025739 · Cited by 15
Adversarial Patch Attacks and Defences in Vision-Based Tasks: A SurveyPreprint
Sharma, A., Bian, Y., Munz, P., & Narayan, A. (2022) · doi:10.36227/techrxiv.20085902 · Cited by 15
LSTM-Based Neural Network to Recognize Human Activities Using Deep Learning Techniques
Sabbu, S., & Ganesan, V. (2022) · Applied Computational Intelligence and Soft Computing · doi:10.1155/2022/1681096 · Cited by 14
“(e architecture was a compact shallow model, and convolutional layers have been applied to the inertial signal spectral domain [7].”
MethodKo, J. E., Kim, S., Sul, J. H., & Kim, S. (2025) · Sensors · doi:10.3390/s25041184 · Cited by 14
“Previous studies [32,42,44,48] converted x, y, and z axes data into separate images for use.”
Yang, G., Zhang, L., Bu, C., Wang, S., Wu, H., & Song, A. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3321639 · Cited by 14
“Implementation details: Following previous popular literatures [14], [39], we apply an optimizer of stochastic gradient descent (SGD) with a momentum parameter of 0.9 and a weight decay of 0.01 to train parameters, where the batch size is constantly set to 512.”
MethodInfluentialSelf-citationSimilarity Embedding Networks for Robust Human Activity Recognition
Li, C., Tong, C. L., Niu, D., Jiang, B., Zuo, X. M., Cheng, L., Xiong, J., & Yang, J. (2021) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3448021 · Cited by 14
“There are many publicly available datasets that are widely adopted by recent studies on HAR [11, 28, 39].”
BackgroundKumar, P., & Chauhan, S. (2021) · CCF Transactions on Pervasive Computing and Interaction · doi:10.1007/s42486-021-00063-5 · Cited by 13
RetractedWaveHAR: Learning Wavelet Representation for Wearable Human Activity Recognition
Yao, M., Zhang, L., Cheng, D., Wu, H., & Song, A. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3571487 · Cited by 13
“To ensure the objectivity of the experimental results [35], we used a window length of 512 data points and a 10% overlap rate to apply sliding window technology when processing the data.”
A Contrastive Learning-Enhanced Residual Network for Predicting Epileptic Seizures Using EEG Signals
Qi, L., Yuan, S., Li, F., Shang, J., Wang, J., & Wang, S. (2025) · International Journal of Neural Systems · doi:10.1142/s0129065725500509 · Cited by 13
Tsokov, S., Lazarova, M. K., & Aleksieva-Petrova, A. (2021) · Indian Journal of Computer Science and Engineering · doi:10.21817/indjcse/2021/v12i2/211202145 · Cited by 12
Exploratory Analysis of Smartphone Sensor Data for Human Activity Recognition
Islam, S. M. M., & Talukder, K. H. (2023) · IEEE Access · doi:10.1109/access.2023.3314651 · Cited by 12
Jangir, M. K., & Singh, K. (2019) · Journal of Discrete Mathematical Sciences and Cryptography · doi:10.1080/09720529.2019.1696552 · Cited by 11
“ActiveMiles [17] is a completely free phone app in a smartphone that allows tracking daily human activity in an uncontrolled environment.”
BackgroundA Review of Wearable Sensor-based Human Activity Recognition using Deep Learning
Zhu, Y., & Mo, L. (2022) · International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence (ICSMD) · doi:10.1109/icsmd57530.2022.10058422 · Cited by 11
“al.[40] LG Nexus 5 smartphone Any placement ActiveMiles”
BackgroundDecision-Making under Uncertainty for the Deployment of Future Hyperconnected Networks: A Survey
Mejía, N. A., Santos‐Boada, G., & Amazonas, J. R. D. A. (2021) · Sensors · doi:10.3390/s21113791 · Cited by 12
Gopinathan, N., & Shanmugam, P. K. (2024) · IEEE Access · doi:10.1109/access.2024.3495018 · Cited by 10
“Attention mechanisms focus on relevant sections of the dataset and often struggle with quadratic complexity and local perception, which can hinder performance in time series owing to noise and non-stationarity (variability) [98].”
From Vision to Motion: Translating Large-Scale Knowledge for Data-Scarce IMU Applications
Yoon, H., Cha, H., Nguyen, H. C., Gong, T., & Lee, S.-J. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3556998 · Cited by 10
“To harness data in all axes for no information loss, we map the x, y, and z axes to the RGB channels to generate a single image, which was shown to be effective in IMU sensing tasks [16], [18].”
InfluentialHuman action recognition based on discriminant body regions selection
Mliki, H., Zaafouri, R., & Hammami, M. (2018) · Signal Image and Video Processing · doi:10.1007/s11760-017-1227-z · Cited by 10
Aprendizado Profundo em Redes Desafiadoras: Conceitos e Aplicações
Bochie, K., Gilbert, M. S., Gantert, L., Barbosa, M., Medeiros, D. S. V., & Campista, M. E. M. (2020) · doi:10.5753/sbc.5033.7.4 · Cited by 10
An efficient CNN model for transportation mode sensing
Tambi, R., Li, P., & Yang, J. K. (2018) · ACM Conference on Embedded Networked Sensor Systems (SenSys) · doi:10.1145/3274783.3275160 · Cited by 10
Hsu, Y.-C., Wang, H., Zhao, Y., Chen, F., & Tsui, K.-L. (2021) · Journal of Medical Internet Research · doi:10.2196/30135 · Cited by 9
Applications of Human Activity Recognition in Different Fields: A Review
Daga, Y., & Meena, S. (2022) · IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON) · doi:10.1109/upcon56432.2022.9986408 · Cited by 9
A framework of transportation mode detection for people with mobility disability
Heo, J., Hwang, S., Moon, J., You, J., Kim, H., Cha, J., & Kim, K. K. (2024) · Journal of Intelligent Transportation Systems · doi:10.1080/15472450.2024.2329901 · Cited by 9
“In addition, the option for data acquisition position in the current application will allow researchers to determine the effect of the position in TMD, as suggested in previous studies (Liang & Wang, 2017; Ravi et al., 2016; Tambi et al., 2018).”
Bernardo, J. B. L., Taparugssanagorn, A., Miyazaki, H., Pati, B. M., & Thapa, U. (2024) · Applied Sciences · doi:10.3390/app142210461 · Cited by 9
Jiang, A., Nacenta, M. A., & Ye, J. (2022) · Visual Informatics · doi:10.1016/j.visinf.2022.05.001 · Cited by 9
“Compared with using a 1D kernel, their 2D kernel method can obtain distinguishable features from multiple sensors; e.g., accelerometers, gyroscopes and magnetometers, and get better performance on common human activity recognition tasks (Jiang and Yin, 2015; Ravi et al., 2016).”
MethodHuman Body Posture Recognition Approaches: A Review
Ali, M. A., Hussain, A. J., & Sadiq, A. T. (2022) · ARO-The Scientific Journal of Koya University · doi:10.14500/aro.10930 · Cited by 9
“However, one limitation of these systems is that they are restricted to the classification activities for which they were developed (Ravi, et al., 2016).”
BackgroundPersonal Health Indicators by Deep Learning of Smart Phone Sensor Data
Li, H., & Trocan, M. (2017) · IEEE International Conference on Cybernetics (CYBCONF) · doi:10.1109/cybconf.2017.7985800 · Cited by 9
“Daniele Ravi brings forward a CNN based human activity recognition method implemented on low-power devices including smartphones [18].”
MethodMemiş, G., & Sert, M. (2019) · IEEE Sensors Journal · doi:10.1109/jsen.2019.2916393 · Cited by 9
“Performances of the proposed model and other state-of-the-art methods on WISDM, Skoda, Daphnet, UCI HAR datasets [19], [23], [36]–[38].”
MethodThe Effect of Axis-Wise Triaxial Acceleration Data Fusion in CNN-Based Human Activity Recognition
Han, X., Ye, J., Luo, J., & Zhou, H. (2020) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2018edp7409 · Cited by 9
Real-time Recognition of Smartphone User Behavior Based on Prophet AlgorithmsPreprint
Mi, C., Xu, R., & Lin, C. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.08997 · Cited by 9
“In order to reduce energy consumption without decreasing recognition accuracy [25][26] , a good observation timing mechanism still needs to be designed.”
BackgroundEgoCHARM: Resource-Efficient Hierarchical Activity Recognition using an Egocentric IMU SensorPreprint
Padmanabha, A., Govindarajan, S., Kim, H., Ortiz, S., Rajan, R., Senkal, D., & Kadetotad, S. (2025) · arXiv.org · doi:10.48550/arxiv.2504.17735 · Cited by 6
“However, performance with decision trees is limited and thus, many recent works have explored neural network architectures to address this gap while still considering compute and memory constraints on the edge [30, 35, 36].”
Madaoui, L., Amira, A., Talha, M. K., Kerdjidj, O., & Himeur, Y. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.107068 · Cited by 8
Yao, M., Cheng, D., Zhang, L., Liu, L., Song, S. J., Wu, H., & Song, A. (2024) · Applied Soft Computing · doi:10.1016/j.asoc.2024.112444 · Cited by 8
Mekruksavanich, S., Tancharoen, D., & Jitpattanakul, A. (2023) · doi:10.1109/incit60207.2023.10413113 · Cited by 8
Deep Learning Shared Bandpass Filters for Resource-Constrained Human Activity Recognition
Stuart, M., & Manic, M. (2021) · IEEE Access · doi:10.1109/access.2021.3064031 · Cited by 8
“To reduce processing overhead, some efforts use general-purpose model reduction techniques [14] or avoid raw data entirely [15].”
MethodSamir, M. A., Mohamed, Z. A., Hussein, M. A. A., & Atia, A. (2021) · IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) · doi:10.1109/uemcon53757.2021.9666662 · Cited by 8
Activity Recognition from Sensor Fusion on Fireman's Helmet
Hackett, S., Cai, Y., & Siegel, M. (2019) · International Congress on Image and Signal Processing (CISP) · doi:10.1109/cisp-bmei48845.2019.8966005 · Cited by 8
“Recently, there has been an increasing popularity in Deep Learning as a classification tool in the area of activity recognition [8].”
BackgroundImproving Resource Efficiency of Deep Activity Recognition via Redundancy Reduction
Leite, C. F. S., & Xiao, Y. (2020) · IEEE Workshop on Mobile Computing Systems and Applications (WMCSA) · doi:10.1145/3376897.3377859 · Cited by 7
“Generic compression methods [2, 3, 5, 8, 11] have been extensively used to make deep models meet the memory, computation, and energy consumption requirements of small devices by reducing the redundancy in the parameters of the neural network.”
MethodA Hybrid Deep Neural Network for Human Activity Recognition based on IoT Sensors
Benhaili, Z., Balouki, Y., & Moumoun, L. (2021) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2021.0121129 · Cited by 7
“It improves accuracy over ensemble learning method [40], and slightly above the model proposed by Ravi et al [41] on the basis of shallow CNN architecture.”
MethodErdelić, M., Carić, T., Erdelić, T., & Tišljarić, L. (2022) · Sustainability · doi:10.3390/su14052756 · Cited by 7
Deep learning approaches for human activity recognition using wearable technology
Janković, M. M., Savić, A. M., Novicic, M. M., & Popović, M. (2018) · Medicinski podmladak · doi:10.5937/mp69-18039 · Cited by 7
The benefit of the doubt: Uncertainty aware sensing for edge computing platforms
Qendro, L., Chauhan, J., Ramos, A. G. C. P., & Mascolo, C. (2021) · arXiv · doi:10.48550/arxiv.2102.05956 · Cited by 6
“Numerous works have investigated the use of deep neural networks for human activity recognition (HAR) [26, 29, 40, 43, 44] and audio sensing [7, 11, 21, 22, 33].”
BackgroundMulti-scale Context-Aware Networks Based on Fragment Association for Human Activity Recognition
Wang, Z., Liu, H., Zhao, B., Shen, Q., Li, M., Que, N., Yan, M., & Xin, J. (2024) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2024/351 · Cited by 6
“Ravi et al.[Ravi et al. , 2016] combined spectrogram features with a single CNN layer and two fully connected layers for HAR, demonstrating the feasibility of real-time applications through evaluations on four benchmark datasets.”
BackgroundLong kernel distillation in human activity recognition
Yao, M., Cheng, D., Zhang, L., Li, W., Wu, H., & Song, A. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113397 · Cited by 6
Deep learning-based badminton action recognition and quality assessment
Li, Y., Feng, Y., Wang, X., & Lu, G. (2025) · Intelligent Data Analysis · doi:10.1177/1088467x251353444 · Cited by 6
A Deep Learning based Human Activity Recognition Model using Long Short Term Memory Networks
Durga, K. M. L., Jyotsna, P., & Kumar, G. K. (2022) · International Conference on Sustainable Computing and Data Communication Systems (ICSCDS) · doi:10.1109/icscds53736.2022.9760794 · Cited by 6
Explaining Human Activities Instances Using Deep Learning Classifiers
Arrotta, L., Civitarese, G., Fiori, M., & Bettini, C. (2022) · IEEE International Conference on Data Science and Advanced Analytics (DSAA) · doi:10.1109/dsaa54385.2022.10032345 · Cited by 6
“The white segments in the first component indicate the occurrence of the following semantic states: DynamicHandManipulations[1, 6], SittingOnDiningChair[1, 6] DynamicHandManipulations[13, 16], InTheDiningRoom[1, 12], and InTheKitchen[13, 16].”
BackgroundZebin, T., Balaban, E., Ozanyan, K. B., Casson, A. J., & Peek, N. B. (2019) · IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI ...) · doi:10.1109/bhi.2019.8834480 · Cited by 6
“Models such as Convolutional Neural Networks (CNN) and RNN employ a data driven approach to learning discriminative features from raw sensor data to infer complex, sequential, and contextual information in a hierarchical manner [11].”
MethodKhataeipour, S. J., Anaraki, J. R., Bozorgi, A., Rayner, M., Basset, F. A., & Fuller, D. (2022) · BMJ Open Sport & Exercise Medicine · doi:10.1136/bmjsem-2021-001242 · Cited by 6
Edge-AI on Wearable Devices: Myocardial Infarction Detection with Spectrogram and 1D-CNN
Gragnaniello, M., Balbi, F., Martellotta, G., Borghese, A., Marrazzo, V. R., Maresca, L., Breglio, G., Irace, A., & Riccio, M. (2024) · IEEE Mediterranean Electrotechnical Conference (MELECON) · doi:10.1109/melecon56669.2024.10608624 · Cited by 5
Evolving 1D Convolutional Neural Networks for Human Activity Recognition
Tsokov, S., Lazarova, M. K., & Aleksieva-Petrova, A. (2021) · International Conference on Computer Systems and Technologies (CompSysTech) · doi:10.1145/3472410.3472413 · Cited by 5
Invited Paper: Hierarchical Activity Recognition with Smartwatch IMU
Alevizaki, A., Pham, N., & Trigoni, N. (2023) · International Conference on Distributed Computing and Networking (ICDCN) · doi:10.1145/3571306.3571390 · Cited by 5
“In principle, the fusion of a multitude of hand-engineered and latent features is often advised in classification tasks to assist the learning process in deep architectures [15, 28].”
Wavelet-domain human activity recognition utilizing convolutional neural networks
Tavakkoli, M., Nazerfard, E., & Amirmazlaghani, M. (2023) · Journal of Ambient Intelligence and Smart Environments · doi:10.3233/ais-230174 · Cited by 5
“HAR data is challenging because it is noisy and contains a large amount of variability due to individual differences in the way people perform activities [22].”
Metadata
- Authors
- Ravi, D., Wong, C., Lo, B., & Yang, G.-Z.
- Affiliations
- Imperial College London (GB)
- Venue
- 2016 IEEE 13th International Conference on Wearable and Implantable Body Sensor Networks (BSN)
- Publisher
- IEEE
- Published
- 2016-06
- Pages
- 71-76
- ISSN
- 2376-8886 (ISSN-L)
- OpenAlex
- W2495920728
- Cited by
- 269
- References
- 22
- Citations by year
- 2026: 11 · 2025: 22 · 2024: 23 · 2023: 27 · 2022: 44 · 2021: 38 · 2020: 36 · 2019: 30 · 2018: 25 · 2017: 11 · 2016: 2
- FWCI
- 7.87
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- Wearable and Implantable Body Sensor Networks (publishedVersion)
- Spiral (Imperial College London) (submittedVersion)
- Funders
- Engineering and Physical Sciences Research Council (EP/H009744/1, EP/L014149/1, EP/E044298/1)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Physical Sciences › Engineering › Biomedical Engineering › Gait Recognition and Analysis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Human Pose and Action Recognition
- Keywords
- human activity recognition, inertial data, wearable devices, real-time classification, temporal convolution, resource-efficient deep learning, mobile devices
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.