Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition
Open accessThe DOI matches its Crossref record; 48 of the 49 references it cites were found in a registry or an index; 8 of the 23 sources its claims cite were checked for support.
Abstract
Introduces DeepConvLSTM, a deep learning architecture combining convolutional and LSTM recurrent units for multimodal wearable activity recognition, outperforming non-recurrent deep networks and handcrafted feature baselines by 4% to 9%.
Overview
CoChat Summary
TL;DR
The paper proposes DeepConvLSTM, a deep framework combining four convolutional layers with two LSTM recurrent layers for wearable activity recognition from raw multimodal sensor data. It reports improved F1 scores over a baseline CNN and prior results on the OPPORTUNITY and Skoda datasets, including 0.915 F1 for OPPORTUNITY gestures with the Null class and 0.958 F1 on Skoda.
Key findings
- On OPPORTUNITY, DeepConvLSTM achieved F1 scores of 0.930 for modes of locomotion without Null, 0.895 for modes with Null, 0.866 for gestures without Null, and 0.915 for gestures with Null; the baseline CNN scored 0.912, 0.878, 0.783, and 0.883, respectively. The paper reports a 6% average improvement over the best OPPORTUNITY challenge submissions and more than 9% improvement on gesture recognition without Null.
- On Skoda, DeepConvLSTM reached 0.958 F1, compared with 0.884 for the baseline CNN, 0.893 for CNN [43], and 0.861 for CNN [23], improving the best reported result by 6%.
- Multimodal fusion improved gesture recognition on OPPORTUNITY without Null: 0.689 F1 with accelerometers only, 0.611 with gyroscopes only, 0.745 with accelerometers + gyroscopes, 0.839 with accelerometers + gyroscopes + magnetic sensors, and 0.864 with the full OPPORTUNITY sensor set.
- DeepConvLSTM had 999,122 parameters versus 7,445,458 for the baseline CNN, a 600% increase for the baseline; training took 340.3 min versus 282.2 min, and classification took 6.68 s versus 5.43 s, with DeepConvLSTM classifying almost 15 min of data per second and achieving 900× real-time on the GPU used.
Methods
- DeepConvLSTM uses four convolutional layers with 64 feature maps each, followed by two dense LSTM layers with 128 cells and a softmax output; the baseline CNN shares the convolutional layers but uses two non-recurrent dense layers with 128 units. Models were trained supervised with cross-entropy loss, mini-batch gradient descent with RMSProp, learning rate 10e-3, decay factor 0.9, mini-batches of 100 segments, and dropout p=0.5 on dense-layer inputs.
- Inputs were sliding windows of 500 ms with 250 ms step size, using raw sensor data with minimal preprocessing: linear interpolation for missing values and per-channel normalization to [0,1]. Evaluation used weighted F1 to address class imbalance.
- Experiments used the OPPORTUNITY dataset (on-body sensors, 113 channels, 30 Hz; Task A: 5-class modes of locomotion/postures; Task B: 18-class sporadic gestures) and the Skoda dataset (one subject, 10 right-arm 3D accelerometers, decimated to 30 Hz, 10 manipulative gestures, about 3 h and about 70 repetitions per gesture). Analyses included comparison to baseline CNN and published results, confusion matrices, sensor-modality fusion, and hyperparameter evaluations of sequence length and number of convolutional layers.
Limitations
- The paper states that DeepConvLSTM has not yet been implemented on a wearable device; the GPU used clearly outperforms current high-end wearable computational power, and training is best envisioned server-side. It also states that pooling was omitted because the sliding-window setup limits downsampling, though pooling could be useful without that requirement, and that training time increases with the number of layers, creating a performance–computational budget trade-off.
- Evident from the text: Skoda evaluation uses one subject and only right-arm sensors, and OPPORTUNITY testing is limited to ADL4 and ADL5 for Subjects 2 and 3, which constrains generalizability. The text does not report cross-validation or statistical significance testing.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures



![Figure 4 Placement of on-body sensors used in the OPPORTUNITYdataset (left: inertial measurements units; right: 3-axis accelerometers) [7].](https://figures.cochat.ai/paper-figures/27c0e4b39d104fe2b5315587d736fd2d/deep-convolutional-and-lstm-recurrent-neural-networks-for-multimodal-wearable-figure-4.jpg)




Verification
Claims the paper makes
5 Valid · 5 Not valid · 13 Not checked · 99 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“The choice of the number of recurrent layers is made following the results presented in [33], where the authors showed that a depth of at least two recurrent layers is beneficial when processing sequential data.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.1. DeepConvLSTM
What the sources say
- Supportedcites [33] Karpathy A. Johnson J. Li F.F. Visualizing and understanding recurrent networks Cornell Univ. Lab. 2015 arXiv:1506.02078 source
Section 4.1 compares recurrent networks with one, two, and three layers on two sequence datasets and states that a depth of at least two is consistently beneficial. This directly supports the attributed finding, within the paper’s tested settings. (LLM, conf=0.94)
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“This seems to indicate that although deep learning techniques are often employed with large amounts of data, (e.g., millions of frames in computer vision [22]), they may actually be applicable to problem domains where acquiring annotated data is very costly, such as in supervised activity recognition.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 5. Results and Discussion > 5.4. Discussion
What the sources say
- Supportedcites [22] Pigou L. Oord A.V.D. Dieleman S. van Herreweghe M. Dambre J. Beyond Temporal Pooling: Recurrence and Temporal Convolutions for Gesture Recognition in Video arXiv Preprint 2015 arXiv:1506.01911 source
The citation is attached to the example of large-scale computer-vision data, not to the citing authors’ conclusion about costly annotation. Section 4.1 (page unavailable) reports 1,720,800 labeled video frames in the gesture-recognition dataset, supporting the cited example of data on the order of millions of frames. (LLM, conf=0.86)
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The baseline CNN also offers better results than the OPPORTUNITY submissions in the recognition of models of locomotion.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 5. Results and Discussion > 5.1. Performance Comparison
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“From the results, we can also see how the model is able to learn from signals obtained from accelerometers, gyroscopes and magnetometers, fusing them without requiring any specific preprocessing.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 6. Conclusions
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“(iv) it is applicable even if the gestures are longer than the observation window.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 5. Results and Discussion > 5.4. Discussion
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
84 more claims
Create a free account to see all 122 claims and to verify the 2 references still open.
- It may state two separate claims as one. 14
- Probably not a substantial claim. 23
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?
The cited source doesn’t support the claim as written.
In the citing paper
“However, these complex network framework suffered from low ef ciency and can hardly meet real-time requirements in practice applications.”
From InnoHAR: A Deep Neural Network for Complex Human Activity Recognition (2019) · cites this paper as [19]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe cited source doesn’t support the claim as written.
In the citing paper
“However, due to the lack of filtration techniques applied to multimodal data, the system’s performance was low.”
From A Multimodal IoT-Based Locomotion Classification System Using Features Engineering and Recursive Neural Network (2023) · cites this paper as [15]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“and yield better performances than pure CNNs or LSTM networks [12]”
From Comparison of Feature Learning Methods for Human Activity Recognition Using Wearable Sensors (2018) · cites this paper as [12]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The proposed method outperforms other methods, such as those based on deep believe networks (DBNs) [10], SVM [10], and CNNs [13].”
From Deep Recurrent Neural Networks for Human Activity Recognition (2017) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The popular DeepConvLSTM model [7] can only afford four, moderately-sized convolution layers to learn effective data representations, and two LSTM layers for sequence modeling.”
From Complex Deep Neural Networks from Large Scale Virtual IMU Data for Effective Human Activity Recognition Using Wearables (2021) · cites this paper as [7]
Experimental — accuracy not yet measured
20 more claims about this paper
Create a free account to see every claim other papers make about this one.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture
“The length of the kernels is the same for every convolutional layer, being defined as Pl=5,∀l=2,⋯,5.”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 4. Experimental Setup > 4.2. Performance Measure
“Following the OPPORTUNITY challenge experimental setup, the length of the window is 500 ms, with a step size of 250 ms.”
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| Total | 996,800 +(128×nc)+nc | 996,800 +(128×nc)+nc | 7,443,136+(128×nc)+nc | 7,443,136+(128×nc)+nc |”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.3. Model Implementation and Training
Reported 996800 · recomputed 1037888
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| 6 | Wai,Waf,Wac,Wao: 7232×128 | 942,592 | W: 57,856×128 | 7,405,696 |”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.3. Model Implementation and Training
Reported 942592 · recomputed 3702784
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“| 7 | Wai,Waf,Wac,Wao: 128×128 | 33,280 | W: 128×128 | 16,512 |”
§ Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition > 3. Architecture > 3.3. Model Implementation and Training
Reported 33280 · recomputed 131072
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Going deeper with convolutions
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298594 · Cited by 48,562 · cited 1×
“Other publications showed much deeper networks, such as “GoogLeNet”, which is a 27-layer deep neural network applied to image classification [45].”
BackgroundReducing the dimensionality of data with neural networks
Hinton, G. E., & Salakhutdinov, R. R. (2006) · Science · doi:10.1126/science.1127647 · Cited by 20,928
Speech recognition with deep recurrent neural networks
Graves, A., Mohamed, A.-R., & Hinton, G. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6638947 · Cited by 8,893 · cited 1×
“Networks using LSTM cells have offered better performance than standard recurrent units in speech recognition, where they gave state-of-the-art results in phoneme recognition [20].”
BackgroundConvolutional networks for images, speech, and time series
LeCun, Y., & Bengio, Y. (1998) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.5555/303568.303704 · Cited by 4,231
The class imbalance problem: A systematic study1
Japkowicz, N., & Stephen, S. (2002) · Intelligent Data Analysis · doi:10.3233/ida-2002-6504 · Cited by 3,295
DeepPose: Human Pose Estimation via Deep Neural Networks
Toshev, A., & Szegedy, C. (2014) · 2014 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2014.214 · Cited by 3,185 · cited 1×
“Deep CNNs have had a major impact on fields, like content recommendation [24], speech recognition [25] and in computer vision [26–28], where they have become the de facto standard approach.”
BackgroundOverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks
Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., & LeCun, Y. (2013) · arXiv · doi:10.48550/arxiv.1312.6229 · Cited by 2,680 · cited 1×
“Deep CNNs have had a major impact on fields, like content recommendation [24], speech recognition [25] and in computer vision [26–28], where they have become the de facto standard approach.”
BackgroundLee, H., Grosse, R., Ranganath, R., & Ng, A. Y. (2009) · Proceedings of the 26th Annual International Conference on Machine Learning · doi:10.1145/1553374.1553453 · Cited by 2,412 · cited 1×
“Deep learning techniques have outperformed many conventional methods in computer vision [13] and audio classification [14].”
MethodA review of wearable sensors and systems with application in rehabilitation
Patel, S., Park, H., Bonato, P., Chan, L., & Rodgers, M. (2012) · Journal of NeuroEngineering and Rehabilitation · doi:10.1186/1743-0003-9-21 · Cited by 2,396 · cited 1×
“, from simple hand gestures to complex activities, such as “cooking a meal”) and the context in which they occur from sensor data is at the core of smart assistive technologies, such as in smart homes [1], in rehabilitation [2], in health support [3,4], in skill assessment [5] or in industrial settings [6].”
Background
Research works printed but not linked: 1
- Gers F.A. Schraudolph N.N. Schmidhuber J. Learning precise timing with LSTM recurrent networks J. Mach. Learn. Res. 2003 3 115 143
Cited by
Items 2,101–2,200 of 2,807
Kaur, D., Pauls, A., & Carvalho, G. H. S. (2025) · IEEE Annual International Computer Software and Applications Conference (COMPSAC) · doi:10.1109/compsac65507.2025.00186 · Cited by 1
Discriminating Between Marijuana and Alcohol Gait Impairments Using Tile CNN With TICA Pooling
Li, R., Uche, S. C., Agu, E., Grimone, K., Herman, D. S., Metrik, J., Abrantes, A. M., & Stein, M. D. (2025) · IEEE Open Journal of Engineering in Medicine and Biology · doi:10.1109/ojemb.2025.3607556 · Cited by 1
“This demonstrates the feasibility of utilizing deep learning methods for prediction of various impairment substances such asrecreationaldrugs,prescriptionmedicationsandclassification ofsymptomsofuserimpairment[67].”
White, M. (2021) · Cronfa (Swansea University) · doi:10.23889/suthesis.58286 · Cited by 2
Human Activity Recognition with Deep Learning: Overview, Challenges & Possibilities
Kumar, P. (2021) · doi:10.20944/preprints202102.0349.v1 · Cited by 12
RetractedHuman Behavior Recognition Based on Millimeter Wave Radar
Chen, Y., Lin, H., Den, P., et al. (2022) · Lecture notes in electrical engineering · doi:10.1007/978-981-19-3927-3_79 · Cited by 1
Csizmadia, G., Daróczy, B., Ferdinandy, B., et al. (2022) · Research Square · doi:10.21203/rs.3.rs-2185125/v1 · Cited by 1
XAttentionHAR Ensemble: Leveraging Cross-Modal Attention for Enhanced Activity Recognition
Sahni, S., Jain, S., & Saritha, S. K. (2024) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001424500265 · Cited by 1
Bazarbekov, I., Almisreb, A. A., Ipalakova, M., Медеуовна, Б. М., & Daineko, Y. A. (2026) · Sensors · doi:10.3390/s26010298 · Cited by 2
“For example, Ordóñez and Roggen (2016) combined CNNs and LSTMs to analyze data from wearable sensors [23].”
Reduction of the Position Bias via Multi-level Learning for Activity Recognition
Osmani, A., & Hamidi, M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-05936-0_23 · Cited by 4
Inertial measurement unit based human action recognition for soft-robotic exoskeletons
Kuschan, J., Burgdorff, M., Filaretov, H., & Krüger, J. (2021) · Proceedings of the Estonian Academy of Sciences · doi:10.3176/proc.2021.4.16 · Cited by 1
Zheng, C., Kasihmuddin, M. S. M., Gao, Y., et al. (2024) · doi:10.1145/3685767.3685784 · Cited by 1
“Ordóñez and Roggen [4] introduced a deep convolutional neural network that leverages data from multiple wearable sensors to recognize a wide range of activities with high accuracy.”
Using Gated Recurrent Units Models for Early Prediction of Sepsis in The Intensive Care UnitPreprint
Zhao, X., & Qu, B. (2021) · Research Square · doi:10.21203/rs.3.rs-259370/v1 · Cited by 1
“Secondly, there exists a specialized gate to determine what the cell states forget and what part of the cell state is written to, which merges the function of different gates in the LSTM structure[22, 37, 38].”
BackgroundMMST-GCN+: A Multi-Modal ST-GCN+ for Computer-Aided Education by Introducing a Sensor Skeleton
Zhao, Z., Song, A., Xiong, Q., Zheng, S., & Guo, J. (2023) · International Conference on Image, Vision and Computing (ICIVC) · doi:10.1109/icivc58118.2023.10270415 · Cited by 2
“For instance, Reyes-Ortiz et al.[21] designed an architecture to classify body action using smartphone and a collection of inertial sensors, which addresses issues the transition problems between known activities and unknown activities; Liu et al.[22] presented a time-representation pattern to capture sequential, inte…”
BackgroundFrom Sensing to Segmentation: Transformer-Based Worker Activity Recognition for Industrial Assembly
Sbaragli, A., & Nardello, M. (2025) · IEEE Access · doi:10.1109/access.2025.3641299 · Cited by 2
“Although the F1-score is 4% lower than the CNN-based results reported in [25], the Transformer converges in 7 minutes and achieves performance suitable for industrial deployment.”
Tran, T., Chang, L., & Lum, P. S. (2025) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2025.1547127 · Cited by 1
Haresamudram, H., Tang, C. I., Suh, S., Lukowicz, P., & Plötz, T. (2023) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3594739.3605102 · Cited by 3
“The advent of Deep Learning (DL) based methods in the general machine learning community was followed by its adoption in the Ubicomp community, leveraging the presence of annotated datasets to learn features that are useful for recognizing activities [3, 8].”
Causal-Inspired Fourier Representation Learning for Wearable IMUs and Egocentric Action Recognition
He, C., Xu, X., Wang, Z., Song, J., Shen, F., & Shen, H. T. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3653125 · Cited by 4
“…recognition task aims to identify human actions by leveraging diverse multi-modal data sources such as skeleton information [6], [7], [15], [16], [17], [25], [26], color images [27], [28], depth maps [6], [10], [29], audio [8], [9], point cloud data [30], and inertial data [13], [18], [19], [31].”
Few-Shot Learning for Detecting Affective States from Keyboard and Mouse Data
Gautam, R., Lin, B., & Yang, M. (2024) · IEEE International Conference on Data Science and Advanced Analytics (DSAA) · doi:10.1109/dsaa61799.2024.10722795 · Cited by 1
“Deep learning techniques, particularly Convolutional Neural Networks (CNNs), have been extensively explored for classification of sequential data [23] [24] [25].”
BackgroundClassification of Covid-19 High-Risk Patients Using Machine Learning
Kothiyal, R., Upadhyay, S., Gola, K. K., et al. (2023) · International Conference on System Modeling & Advancement in Research Trends (SMART) · doi:10.1109/smart59791.2023.10428283 · Cited by 1
Hand Posture Detection of Smartphone Users Using LSTM Networks
Tan, S., Ng, H., Ooi, B.-Y., et al. (2019) · Lecture notes in electrical engineering · doi:10.1007/978-981-13-6447-1_3 · Cited by 1
Sigcha, L. (2021) · UPM Digital Archive (Technical University of Madrid) · doi:10.20868/upm.thesis.68904 · Cited by 4
Human Object Interactivity Detection usingGraph Neural Networks
Harbola, A., Rana, R., Rawat, S. S., Saini, K., Vats, S., & Sharma, V. (2024) · International Conference on Electronics, Computing, Communication and Control Technology (ICECCC) · doi:10.1109/iceccc61767.2024.10593986 · Cited by 1
“NTU-RGBD -120 : This dataset [9] is a subset of the larger NTU-RGBD collection of over 100,000 video clips showing people performing 120 different actions.”
MethodFedFitTech: A Baseline in Federated Learning for Fitness Tracking
Oz, Z., Korde, S., Bock, M., & Laerhoven, K. V. (2025) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3714394.3756332 · Cited by 1
“Over the past decade, Deep Learning (DL) methods have become the dominant approach in inertial-based HAR, consistently outper-forming traditional Machine Learning algorithms [24, 25].”
Sensor-based gait recognition using ensemble network unified with independent subnetworks
Das, S., & Sahoo, P. R. (2025) · IEEE Region 10 Conference (TENCON) · doi:10.1109/tencon66050.2025.11375107 · Cited by 1
“…datasets: WHU-GAIT [13], comprising data from [11] 92.91% 96.78% 99.58% CNN [25] 92.89% 97.02% 99.71% LSTM [26] 91.88% 96.98% 99.46% DeepConv [27] 92.25% 96.80% 99.24% CNN+LSTM [2] 92.51% 96.82% 99.61% CNN fix +LSTM [2] 92.94% 97.04% 99.64% CNN+LSTM fix [2] 93.52% 97.33% 99.75% LSTM & CNN…”
InfluentialDCU team at the NTCIR-15 micro-activity retrieval task
Le, T.-K., Nguyen, M.-D., Tran, L.-D., Ninh, V.-T., Gurrin, C. G., & Healy, G. F. (2020) · Dublin City University Open Access Institutional Repository (Dublin City University) · Cited by 1
Unveiling Human Activity Patterns in Smart Cities Through a CNN-LSTM Approach
Ameur, I., Ameur, M. E. A., Ameur, M., & Degha, H. E. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-82112-7_4 · Cited by 2
Shi, J., Xi, T., Fan, X., Tang, J., & Yu, H. (2024) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2024.3479314 · Cited by 1
“In recent years, deep learning has been increasingly applied to the field of human emotion recognition [28], [29], [30].”
RDPD: Rich Data Helps Poor Data via ImitationPreprint
Hong, S., Xiao, C., Hoang, T. N., et al. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1809.01921 · Cited by 1
Revisiting Backdoor Attacks on Time Series Classification in the Frequency Domain
Huang, Y., Zhang, M., Wang, Z., Li, W., & Yang, M. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714827 · Cited by 12
“In the initial stages, human-crafted features were employed [37, 50].”
Research on Dangerous Driving Behavior Recognition Algorithm Based on LSTM
Huang, Y., Zhang, M., Xue, M., et al. (2025) · SAE technical papers on CD-ROM/SAE technical paper series · doi:10.4271/2025-01-8204 · Cited by 1
A systematic review of deep learning-based models for elderly and human activity recognition
Dalal, M., & Mittal, P. (2025) · Computer Science Review · doi:10.1016/j.cosrev.2025.100879 · Cited by 1
Continual Egocentric Activity Recognition via Foreseeable-Generalized Multi-Modal Representations
He, C., Cheng, S., Qiu, Z., Xu, L., Meng, F., Wu, Q., & Li, H. (2023) · doi:10.36227/techrxiv.24041583.v1 · Cited by 11
A Complex User Activity Recognition Algorithm based on Convolutional Neural Networks
Zhang, J., Luo, H., Chen, S., Zhao, F., & Jiang, M. (2018) · Ubiquitous Positioning, Indoor Navigation, and Location Based Service · doi:10.1109/upinlbs.2018.8559850 · Cited by 1
“The output of a convolution layer can be defined as [23] :”
Introducing and Benchmarking a One-Shot Learning Gesture Recognition Dataset
Kasnesis, P., Chatzigeorgiou, C., Patrikakis, C. Z., & Rangoussi, M. (2021) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-030-72802-1_8 · Cited by 1
A learning behavior classification model based on classroom meta-action sequences
Shou, Z., Yuan, X., Li, D., Mo, J., Zhang, H., Hua, Y., & Wu, Z. (2025) · Scientific Reports · doi:10.1038/s41598-025-06901-1 · Cited by 1
SCAR-Net: Scalable ConvNet for Activity Recognition with Multimodal Sensor Data
Nazi, Z. A. (2020) · Smart innovation, systems and technologies · doi:10.1007/978-981-15-8269-1_5 · Cited by 1
Automated Lubrication Systems Prognostics Using Long - Term Recurrent Convolutional Networks
Warner, C., & Desmet, A. (2018) · IEEE Conference on Prognostics and Health Management · doi:10.1109/icphm.2018.8448900 · Cited by 3
“4) Hybrid Neural Networks: Hybrid neural networks, combining convolutional and LSTM networks, have proven capable at sequential analysis in spatially- and temporallycorrelated 1D [28] and 2D [29] datasets, with better disambiguation between closely-related classes compared to standard convolutional approaches.”
BackgroundSawsen, G., Ouarda, W., Boubaker, H., Khelifa, M. M. B., & Alimi, A. M. (2021) · doi:10.36227/techrxiv.13991198 · Cited by 1
Kojima, R., Legaspi, R., Yoshihara, K., & Wada, S. (2020) · Smart innovation, systems and technologies · doi:10.1007/978-981-15-8269-1_3 · Cited by 2
BackinFocus: Iterative Design and Validation of a Smart Jacket for Robust Posture Sensing
Kalvoda, V. E., Pribbernow, J., Rieger, N., & Wolf, T. M. (2026) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/3772363.3799182
“While early HAR relied on handcrafted features [2, 29], deep learning transformed the field through automatic feature extraction, progressing from MLPs [19] to CNNs for spatial dependencies [8, 31], LSTMs for temporal modeling [22], and hybrid CNN-LSTM architectures [9, 20].”
Sansano‐Sansano, E. (2020) · Tesis Doctorals en Xarxa (Consorci de Serveis Universitaris de Catalunya) · doi:10.6035/14101.2020.180044 · Cited by 4
Niemann, F., Rueda, F. M., Kfari, M. K. A., Nair, N. R., Schauten, D., Kretschmer, V., Lüdtke, S., & Kirchheim, A. (2026) · Sensors · doi:10.3390/s26020739 · Cited by 1
S-LIME: An Explainable SVM-Based Framework for Human Activity Recognition Using LIME
Karim, M. M., Kabir, S., & Fattah, S. A. (2025) · IEEE Region 10 Conference (TENCON) · doi:10.1109/tencon66050.2025.11375496 · Cited by 1
“Convolutional Neural Networks (CNNs) [4], Recurrent Neural Networks (RNNs) [5], and more recently Transformer-based architectures [6] have been successfully applied to HAR, achieving remarkable accuracy improvements on benchmark datasets such as UCI-HAR [7], WISDM [8], and KU-HAR [9].”
Radio Signal Recognition Using Two-Stage Spatiotemporal Network with Bispectral Analysis
Bai, H., Li, S., Yong Qiang, J. I. A., & Xiao, B. (2025) · Sensors · doi:10.3390/s25175449 · Cited by 1
“This approach has proven effective in domains relying on spatiotemporal dependencies, such as acoustic event detection [20], human activity recognition [21], and EEG-based emotion classification [22].”
Machine Learning Applied to Locomotion in Virtual Reality
Sakabe, F. K., Ayres, F. J., & Soares, L. P. (2024) · Symposium on Virtual and Augmented Reality (SVR) · doi:10.1145/3691573.3691592 · Cited by 3
Att-DSConv: Separation of I/Q Signals Based on Self-Attention and Depthwise Separable Convolution
Lin, F., Yu, X., & Li, H. (2024) · 2024 4th International Conference on Computer Systems (ICCS) · doi:10.1109/iccs62594.2024.10795860 · Cited by 1
Towards Putting Deep Learning on the Wrist for Accurate Human Activity Recognition
Bohra, S., Naik, V. S., & Yeligar, V. (2021) · IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops) · doi:10.1109/percomworkshops51409.2021.9430979 · Cited by 2
Human Activity Recognition Implementing CNN-LST
Negrete, P. D. M., Flores, A., Minango, J., & Zambrano, M., V (2024) · Smart innovation, systems and technologies · doi:10.1007/978-3-031-66961-3_53 · Cited by 1
Zohrevand, A., Mirzaei, S., & Sajedi, H. (2025) · International Computer Conference, Computer Society of Iran (CSICC) · doi:10.1109/csicc65765.2025.10967410 · Cited by 1
“The network structure follows recent best practices in deep learning for sensor-based HAR [14, 11].”
EEG-ConvoBLSTM: A novel hybrid model for efficient EEG signal classification
Zhang, L., Zhang, X., Zhang, X., & Yang, Y. (2025) · Review of Scientific Instruments · doi:10.1063/5.0258358 · Cited by 2
Çeçen, M., Rivera, M. E., & Yavuz, M. (2026) · Wind Energy · doi:10.1002/we.70114 · Cited by 1
Li, G., Lim, C. H., Ma, Q., Tang, X., Tew, H. H., Ding, F., & Luo, X. (2024) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc54092.2024.10831884 · Cited by 2
“Among these researchers, DeepConvLSTM proposed by [2] is one of the deep learning architectures that has attracted a lot of attention in the study of HAR.”
Wang, C., Shi, H., Song, B., et al. (2024) · Journal on Computing and Cultural Heritage · doi:10.1145/3703633 · Cited by 1
“The recurrent neural network (RNN) [10], a commonly used method for temporal feature extraction, faces the challenges of gradient disappearance and long-term dependency.”
Game of LLMs: Discovering Structural Constructs in Activities using Large Language Models
Hiremath, S. K., & Plötz, T. (2024) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3675094.3678444 · Cited by 4
“Once identified such components are modeled using sequential procedures like HMMs, LSTMs, or RNNs in general [15, 32].”
Human Action Recognition in Multi-Level Convolutional Temporal Attention Network
Huang, Q., Chen, Z., Li, C., & Qian, W. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889140 · Cited by 1
“Some researchers have found that models combining CNN and LSTM [7] can effectively capture the features and tem-*Corresponding author. poral dependencies of time series data.”
A Context-aware Hierarchical Approach for Activity Recognition Based on Mobile Devices
Zhang, S., Wei, Z., Nie, J., Huang, L., & Li, Z. (2017) · Computer Systems Science and Engineering · Cited by 1
Rare Disease Phenotyping via Self-Supervised Multimodal Learning
Uppalapati, K., Yimenicioglu, B., Eftekhari, A., & Abdulkareem, S. (2025) · ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics (ACM-BCB) · doi:10.1145/3765612.3767304 · Cited by 1
“D medical scans lacking clinical context [31], and sensor fusion methods that fail to incorporate higher-level domain knowledge [26].”
Elderly Exercise Activity Classification Based on Pose Estimation Using CNN-LSTM
Ulya, A. R. A., Yuniarno, E. M., Muhtadin, M., & Purnomo, M. H. (2024) · International Seminar on Intelligent Technology and Its Applications · doi:10.1109/isitia63062.2024.10667962 · Cited by 5
Deep Dissimilarity Measure for Trajectory Analysis
Arfa, R., Yusof, R. B., & Shabanzadeh, P. (2018) · Communications in computer and information science · doi:10.1007/978-981-13-2853-4_11 · Cited by 1
Gowrishankar, R., Leite, C. F. S., Mauranen, H., & Xiao, Y. (2025) · International Conference on Human-Computer Interaction with Mobile Devices and Services Adjunct (MobileHCI Adjunct) · doi:10.1145/3737821.3749576 · Cited by 4
Wang, P., Guo, F., Gu, F., Li, M., & Long, X. (2024) · International Conference on Mobility, Sensing and Networking (MSN) · doi:10.1109/msn63567.2024.00116 · Cited by 2
Human Activity Recognition Based on a Cascade Network Architecture Composed of DFSMN and ACGAN
Xie, J., Liang, X., & Xing, C. (2023) · International Conference on Intelligent Human-Machine Systems and Cybernetics · doi:10.1109/ihmsc58761.2023.00053 · Cited by 1
“To validate the performance of our model, we compared it with other networks using evaluation metrics, such as LSTM-CNN, Attention-Based CNN, and H-LSTM, the comparison results are shown in Table 2.”
MethodInfluentialZhang, J. A., & Ebling, M. R. (2023) · IEEE MIT Undergraduate Research Technology Conference (URTC) · doi:10.1109/urtc60662.2023.10535025 · Cited by 1
Meng, Y., Liu, Z., Li, S., & Sun, T. (2025) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3714394.3756174 · Cited by 1
“This combination has proven effective for multimodal sensor-based activity recognition [22].”
Sun, Q., Li, Z., Wang, K., Fu, Z., Chen, G., Zhang, L., Wu, H., & Song, A. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3694938 · Cited by 2
“Building on this, Ord´o˜nez and Roggen [7] developed DeepConvLSTM, an eight-layer hybrid architecture that uses convolutional layers for local features and LSTMs for long-range dependencies, outperforming purely convolutional backbones for sensor-based HAR.”
Lightweight Transformer in Federated Setting for Human Activity RecognitionPreprint
Raza, A., Tran, K. P., Koehl, L., Li, S., Zeng, X., & Benzaidi, K. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.00244 · Cited by 13
“Recently, by utilizing deep learning, researchers have made substantial progress on high-accuracy HAR classifiers [4, 5, 6].”
BackgroundDiffusion-Driven Inertial Generated Data for Smartphone Location Classification
Cohen, N., Dror, R., & Klein, I. (2025) · International Conference on Indoor Positioning and Indoor Navigation · doi:10.1109/ipin66788.2025.11213024 · Cited by 3
Rotating Stall Detection of Axial Compressors Based on Multi-source Heterogeneous Data Fusion
Li, Y., Quan, F., & Sun, X. (2022) · China Automation Congress (CAC) · doi:10.1109/cac57257.2022.10054901 · Cited by 1
“Especially, the combination of a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN) has been proven effective in multiple time series classification tasks [13]– [15], because they can fit complex systems without requiring additional domain knowledge.”
BackgroundTurdalyuly, M., Zholdassova, A., Turdalykyzy, T., & Doshybekov, A. (2026) · Information · doi:10.3390/info17040368 · Cited by 1
Retrieval-based Annotation of Multi-channel Time-Series Data for HAR
Altermann, E., Rueda, F. M., Rusakov, E., & Fink, G. A. (2022) · IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops) · doi:10.1109/percomworkshops53856.2022.9767355 · Cited by 1
“The networks are trained by minimizing the binary-cross entropy loss using stochastic gradient-descent with RMSProp update rule [30], [17].”
MethodNRBO-AGP: A novel feature selection approach for accurate protein solubility prediction
Elmi, Z., Elmi, S., & Danishvar, S. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.129194 · Cited by 4
A deep learning framework for gait-based frailty classification using inertial measurement units
Amjad, A., Szczęsna, A., Błaszczyszyn, M., Sacha, J., Sacha, M., Feusette, P., Wolański, W., Konieczny, M., Borysiuk, Z., & khan, B. (2026) · PLoS ONE · doi:10.1371/journal.pone.0343402 · Cited by 1
Reddy, N. N. K., Govindu, P. B., Swaminathan, B., & Subramaniyaswamy, V. (2026) · Progress in Artificial Intelligence · doi:10.1007/s13748-026-00436-w · Cited by 1
“Ordóñez and Roggen also discussed physical activity recognition using LSTMs, and heavily relied on what a sequence model may offer in such a context [25].”
Bai, Y., Yang, W., Zhu, B., Zhai, Z., Wang, Z., & Zou, X. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3638891 · Cited by 2
“Conventional “sense-then-transmit” sensors suffer from high data-link power, bandwidth bottlenecks, and sluggish response [1], [3], and their common reliance on deep neural networks [4], [5], [6], [7] further incurs substantial computational complexity and heavy resource demands.”
Human Activity Recognition and Control of Wearable Robots
Chinimilli, P. T. (2018) · Cited by 3
Xu, Z. M., Zhu, Y. C., & Li, S. (2026) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2026.1901746 · Cited by 1
Vsens: Incorporating XR into the Process of Collecting Virtual IMU Data
Liang, F., Min, T., Xia, C., & Sugiura, Y. (2026) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3790110 · Cited by 1
“For the deep-learning baseline (ConvLSTM [47]), both virtual and real signals were resampled to 20 Hz and fed as temporal sequences.”
Human Activity Recognition via a Categorical Kalman Filter Using Inertial Sensors
Pitombeira-Neto, A. R., de França, D. S., Cruz, L. A., & de Macêdo, J. A. D. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3678199
“Noteworthy models have been proposed in the literature, including multilayer perceptrons (MLPs) [23], [24], convolutional NNs (CNNs) [25], [26], [27], deep recurrent NNs (RNNs) [28], [29], NN with attention mechanism [30], and ResNet [31].”
Zhang, J., & Oh, S.-S. (2026) · International Journal of Computational Intelligence Systems · doi:10.1007/s44196-026-01196-0 · Cited by 3
A Fusion Neural Network Incorporating Attention for Sensor-Based Human Activity Recognition
Lu, L., Zhang, C., Cao, K., & Deng, D. (2022) · 2022 3rd International Conference on Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA) · doi:10.1109/cvidliccea56201.2022.9824475 · Cited by 1
A Special issue on Multidisciplinary research to promote the applications in Biological sciences
Manikandan, G. E. D. M. (2022) · International Journal of Life Science and Pharma Research · doi:10.22376/ijpbs/ijlpr/sp22/2022.1-154 · Cited by 1
COMODO: Cross-Modal Video-to-IMU Distillation for Efficient Egocentric Human Activity Recognition
Chen, B., Wongso, W., Li, Z., Khaokaew, Y., Xue, H., & Salim, F. D. (2026) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3810218 · Cited by 1
A multi-scale feature extraction fusion model for human activity recognitionPreprint
Zhang, C., Cao, K., Lu, L., et al. (2022) · Research Square · doi:10.21203/rs.3.rs-1933621/v1 · Cited by 1
Patrick, B., Kanjo, E., & Kaiwartya, O. (2026) · Smart Agricultural Technology · doi:10.1016/j.atech.2026.101986 · Cited by 5
Duan, F., Zhu, T., Chen, L., Ning, H., Liu, C., & Wan, Y. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3543928 · Cited by 1
“Ordóñez and Roggen [18] devised a DeepConvLSTM, which combines convolutional neural networks and a two-layer LSTM to extract temporal and spatial features.”
Am I fit for this physical activity? Neural embedding of physical conditioning from inertial sensorsPreprint
de Aguiar, D. P., & Murai, F. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.12095 · Cited by 1
“DeepConvLSTM [9] is an architecture designed for the HAR task.”
BackgroundMethodInfluentialData-Driven Estimation of Rowing Forces and Power via Long Short-Term Memory Neural Networks
Meyer, L., Schmid, A., & Braun, S. (2024) · Bulletin of pioneering researches of medical and clinical science · doi:10.51847/gjr2udtuzl · Cited by 1
Li, C., Huang, J., & Pei, J. (2024) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls61622.2024.10606643 · Cited by 1
Continuous physical activity recognition for intelligent labour monitoring
Motrenko, A., Simchuk, E., Khairullin, R., Inyakin, A. S., Kashirin, D., & Strijov, V. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-11288-y · Cited by 3
“Application of deep learning for behaviour analysis have been reported in [1, 19, 30, 40].”
BackgroundFedOpenHAR: Federated Multitask Transfer Learning for Sensor-Based Human Activity Recognition
İşgüder, E., & İncel, Ö. D. (2025) · Journal of Computational Biology · doi:10.1089/cmb.2024.0631 · Cited by 8
“Federated learning is mostly employed using a single source of data and then splitting the source into multiple “ slices ” to obtain a scheme where each slice is then dedicated to a participant of the federated learning approach, referred to as “ clients ” (Yang et al., 2019).”
InfluentialZhang, K., Liu, Y., Ferré, M., & Badesa, F. J. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110296 · Cited by 2
Human Activity Recognition Model Capable of Handling Various Input Waveforms
Hasegawa, T. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-6606-5_1 · Cited by 2
Domain Human Recognition Techniques using Deep Learning
Merikapudi, S., Murthy, S. V. N., A, M. S., & Gandhi, R. (2022) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2022.0130684 · Cited by 1
Fruit Picker Activity Recognition with Wearable Sensors and Machine Learning
Dabrowski, J. J., & Rahman, A. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191571 · Cited by 2
“To our knowledge HAR has not been applied to monitor fruit pickers.”
BackgroundMethodInfluentialLearning Sensor Sample-Reweighting for Dynamic Early-Exit Activity Recognition Via Meta Learning
Fu, Z., Zhang, L., Huang, W., Cheng, D., Wu, H., & Song, A. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3524252 · Cited by 5
He, T., Yi, S., Zeng, H., Mei, Z., Ivanov, K., Yang, L., Tang, J., Tang, J., & Mei, Z. (2026) · Neural Computing and Applications · doi:10.1007/s00521-025-11832-z · Cited by 1
Object-Centric Approach to Prediction and Labeling of Manipulation Tasks
Chen, E. H., & Burschka, D. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8462973 · Cited by 1
“[19] investigates the potential of combining CNN with a LSTM for multimodal sensors.”
BackgroundCattle Behavior Classification Based on Integration of Acceleration Sensor and Position Information
Tian, H., Lvwen, H., Mengqun, Z., et al. (2023) · International Conference on Computer Communication and the Internet (ICCCI) · doi:10.1109/iccci59363.2023.10210161 · Cited by 1
“We applied the sensor data augmentation methods mentioned in [26] for the collected sequence data, including scaling, dithering and random rotation.”
MethodA Federated Learning Based Hybrid Deep Learning Framework for Enhanced Human Activity Recognition
Azmoudeh, J., Arghaee, S., Valizadeh, P., Dandani, S., Havangi, I., & Yaghmaee, M. H. (2024) · International Conference on Computer and Knowledge Engineering · doi:10.1109/iccke65377.2024.10874707 · Cited by 2
Metadata
- Authors
- Ordóñez, F., & Roggen, D.
- Affiliations
- University of Sussex (GB)
- Corresponding author
- Ordóñez, Francisco
- Venue
- Sensors
- Publisher
- MDPI AG
- Published
- 2016-01-18
- Volume
- 16
- Issue
- 1
- Pages
- 115
- ISSN
- 1424-8220 (ISSN-L)
- PMC
- PMC4732148
- OpenAlex
- W2270470215
- Cited by
- 2,761
- References
- 49
- Citations by year
- 2026: 284 · 2025: 256 · 2024: 284 · 2023: 270 · 2022: 319 · 2021: 379 · 2020: 347 · 2019: 288 · 2018: 219 · 2017: 93 · 2016: 21
- FWCI
- 140.52
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Sensors (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Sussex Research Online (University of Sussex) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Figshare (submittedVersion, cc-by)
- Funders
- Google Faculty Research Award (Is deep learning useful for wearable activity recognition?)
- EPSRC (EP/N007816/1)
- 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 › Signal Processing › Time Series Analysis and Forecasting
- Keywords
- human activity recognition, wearable sensors, multimodal sensor fusion, convolutional neural networks, LSTM neural network, temporal dynamics, automatic feature extraction, hyperparameter optimization
- MeSH terms
- Machine Learning, Clothing, Humans, Human Activities, Signal Processing, Computer-Assisted, Databases, Factual, Neural Networks, Computer, Monitoring, Ambulatory
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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.