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 1,101–1,200 of 2,808
A Smart Home Simulation Tool to Support the Recognition of Activities of Daily Living
Ho, B., Vogts, D., & Wesson, J. L. (2019) · South African Institute of Computer Scientists and Information Technologists · doi:10.1145/3351108.3351132 · Cited by 19
Sezavar, A., Atta, R., Ghanbari, M., & Fellow, I. L. (2022) · Multimedia Systems · doi:10.1007/s00530-022-00954-2 · Cited by 15
“Using the DTCWT within the proposed model provides a better feature representation for gait recognition and a better recognition accuracy than the other hybrid deep learning networks as in [11, 19–21].”
MethodHuman Action Recognition from Various Data Modalities: A ReviewPreprint
Rahmani, H., Bennamoun, M., & Ke, Q. (2021) · doi:10.36227/techrxiv.13708270.v1 · Cited by 14
Hernandez, V., Kulić, D., & Venture, G. (2020) · Journal of Biomechanics · doi:10.1016/j.jbiomech.2020.109684 · Cited by 14
Understanding Smartwatch Battery Utilization in the Wild
Homayounfar, M., Malekijoo, A., Visuri, A., Dobbins, C., Peltonen, E., Pinsky, E., Teymourian, K., & Rawassizadeh, R. (2020) · Sensors · doi:10.3390/s20133784 · Cited by 15
Context-Aware Complex Human Activity Recognition Using Hybrid Deep Learning Models
Omolaja, A., Otebolaku, A., & Alfoudi, A. S. D. (2022) · Applied Sciences · doi:10.3390/app12189305 · Cited by 21
“As argued by [22,58], we fed the output of the single-layer LSTM network to a fully connected dense layer with a SoftMax activation function.”
MethodEzzeldin, M., Ghoneim, A. S., Abdelhamid, L., & Atia, A. (2024) · Signal Image and Video Processing · doi:10.1007/s11760-024-03552-z · Cited by 22
“Ordóñez and Roggen [8] introduced a DNN framework using DeepConvLSTM for activity classification.”
Motion Pattern Recognition for Indoor Pedestrian Altitude Estimation Based on Inertial Sensor
Guo, Q., Xia, M. Y., Yan, D., Wang, J., Shi, C., Wang, Q., & Li, T. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3355163 · Cited by 15
“Notably, Ordonez and Roggen [28] merged CNNs and LSTMs to train and test on two human activity recognition datasets (OPPORTUNITY and Skoda), revealing the superiority of neural networks over traditional machine learning methods.”
MethodHuman Activity Recognition Method Based on Edge Computing-Assisted and GRU Deep Learning Network
Huang, X., Yuan, Y., Chang, C., Gao, Y., Zheng, C., & Yan, L. (2023) · Applied Sciences · doi:10.3390/app13169059 · Cited by 25
Temporal Action Localization for Inertial-based Human Activity Recognition
Bock, M., Moeller, M., & Laerhoven, K. V. (2024) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3699770 · Cited by 29
“In their original publication Ordóñez and Roggen [35] have found a combination of convolutional and recurrent layers to produce competitive results, with the idea being to model temporal dependencies amongst automatically extracted discriminative features within a sliding window in order to…”
InfluentialUser Authentication by Gait Data from Smartphone Sensors Using Hybrid Deep Learning Network
Cao, Q., Xu, F., & Li, H. (2022) · Mathematics · doi:10.3390/math10132283 · Cited by 16
Liu, J., Wei, Y., Berges, M. E., Bielak, J., Garrett, J. H., & Noh, H. Y. (2019) · Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems · doi:10.1117/12.2513711 · Cited by 15
Kwon, H., Tong, C., Haresamudram, H., Gao, Y., Abowd, G. D., Lane, N. D., & Ploetz, T. (2020) · arXiv · doi:10.48550/arxiv.2006.05675 · Cited by 14
“Method We employ two machine learning models, Random Forest and DeepConvLSTM [54].”
MethodInfluentialSong, Y., Li, Y., Wang, Y., Ye, C., Zhu, C., & Yang, R. (2026) · Advanced Functional Materials · doi:10.1002/adfm.202532014 · Cited by 28
“As model depth or width increases, however, mobile deployment must balance accuracy with compute and memory budgets [186, 191].”
Hasan, M. A., Li, F., Gouverneur, P., Piet, A., & Grzegorzek, M. (2025) · PLoS ONE · doi:10.1371/journal.pone.0315343 · Cited by 22
“We selected the Deep convolutional and LSTM recurrent neural networks (DeepConvL-STM) model proposed by Ordonez et. al [4] as the benchmark model for our study as shown in Fig 4.”
Jiang, X., Chen, R., & Zhu, H. (2021) · APL Materials · doi:10.1063/5.0043842 · Cited by 18
Comparison of Keras Optimizers for Earthquake Signal Classification Based on Deep Neural Networks
Mustika, I. W., Adi, H. N., & Najib, F. (2021) · International Conference on Information and Communications Technology (ICOIACT) · doi:10.1109/icoiact53268.2021.9563990 · Cited by 20
Noreldeen, H. A. A., Huang, K., Wu, G.-W., Zhang, Q., Peng, H., Deng, H., & Chen, W. (2022) · Analytical Chemistry · doi:10.1021/acs.analchem.2c03814 · Cited by 14
3D Dynamic Image Modeling Based on Machine Learning in Film and Television Animation
Wang, Y. (2023) · Journal of Multimedia Information System · doi:10.33851/jmis.2023.10.1.69 · Cited by 15
A deep learning approach for detecting tic disorder using wireless channel information
Barua, A., Dong, C., & Yang, X. (2020) · Transactions on Emerging Telecommunications Technologies · doi:10.1002/ett.3964 · Cited by 14
MICAR: multi-inhabitant context-aware activity recognition in home environments
Arrotta, L., Bettini, C., & Civitarese, G. (2022) · Distributed and Parallel Databases · doi:10.1007/s10619-022-07403-z · Cited by 19
“In our experiments, we implemented a Deep Neural Network which combines convolutional (CNN) and recurrent (LSTM) layers, inspired by the work proposed in [38].”
MethodBian, S., Yuan, S., Rey, V. F., & Lukowicz, P. (2022) · Smart innovation, systems and technologies · doi:10.1007/978-981-19-0361-8_5 · Cited by 16
Titchy: Online Time-Series Compression With Random Access for the Internet of Things
Vestergaard, R., Zhang, Q., Sipos, M., & Lucani, D. E. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3081868 · Cited by 22
Comparing Cross-Subject Performance on Human Activities Recognition Using Learning Models
Yang, Z., Qu, M., Pan, Y., & Huan, R. (2022) · IEEE Access · doi:10.1109/access.2022.3204739 · Cited by 18
“Ordonez et al. [21] proposed a general network framework consisting of a four-layer CNN and a two-layer LSTM for behavior recognition coined as DeepConvLSTM, which achieved a 7.4% and 3.2% perfor-183 mance improvement over the original CNN baseline model 184 in the Skoda [43] and Opportunity…”
MethodInfluentialAdaptive sequencing using nanopores and deep learning of mitochondrial DNA
Danilevsky, A., Polsky, A. L., & Shomron, N. (2022) · Briefings in Bioinformatics · doi:10.1093/bib/bbac251 · Cited by 21
Jung, S., Bong, J. H., Kim, K., & Park, S. (2024) · Frontiers in Bioengineering and Biotechnology · doi:10.3389/fbioe.2023.1272693 · Cited by 18
Yasutomi, A. Y., & Enoki, H. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2974709 · Cited by 14
“With the advance of 1D time series classification technology with DNNs [9]–[11], this generalization can be more effortlessly achieved, further expanding the position estimation capabilities.”
MethodRobust 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
“Increasing the model parameters (e.g., introducing more layers or filters) has been shown to enhance model performance [16], [17].”
BackgroundMethodResultInfluentialA GAN-Based Data Augmentation Approach for Sensor-Based Human Activity Recognition
Chen, W., & Cho, P.-C. (2021) · International Journal of Computer and Communication Engineering · doi:10.17706/ijcce.2021.10.4.75-84 · Cited by 17
“Combining CNNs and RNNs to train a hybrid model can take advantages from both CNN and RNN models and achieve better results in recognizing Volume 10, Number 4, October 2021 various activities [13].”
MethodGOAT: A Generalized Cross-Dataset Activity Recognition Framework with Natural Language Supervision
Miao, S., & Chen, L. (2024) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3699736 · Cited by 27
“To evaluate the effectiveness of cosine similarity loss in our framework, we compared GOAT’s performance with different loss functions across three model architectures: DeepConvLSTM, GOAT-Sup, and GOAT.”
InfluentialMale, J., & Martínez-Hernández, U. (2021) · IEEE International Conference on Industrial Technology (ICIT) · doi:10.1109/icit46573.2021.9453672 · Cited by 16
“Various assembly task actions common to an industrial setting have been investigated previously [12][13], the screwdriver and Allen key movement actions are relatively similar so provide some challenge for a classifier to identify.”
BackgroundHaresamudram, H., Essa, I., & Plötz, T. (2024) · Sensors · doi:10.3390/s24041238 · Cited by 14
“The de facto approach for obtaining features is to compute them via statistical metrics [6,7] and the empirical cumulative distribution function [8], or to learn them directly from the data itself (e.g., via end-to-end training [9–13] or unsupervised learning [14–16]).”
MethodOn-Device Deep Personalization for Robust Activity Data Collection
Mairittha, N., Mairittha, T., & Inoue, S. (2020) · Sensors · doi:10.3390/s21010041 · Cited by 20
Wi-Adaptor: Fine-grained Domain Adaptation in WiFi-based Activity Recognition
Zhang, H., Zhou, Z., & Gong, W. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685725 · Cited by 16
“Many traditional sensing technologies use cameras [4], [15], wearable sensors [5], [16], or radiofrequency identification [6], [17] for recognition.”
BackgroundLakoju, M., Ajienka, N., Khanesar, M. A., Burnap, P., & Branson, D. T. (2021) · Sensors · doi:10.3390/s21154991 · Cited by 16
1D Convolutional Neural Network with Long Short-Term Memory for Human Activity Recognition
Goh, J. X., Lim, K. M., & Lee, C. P. (2021) · IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET) · doi:10.1109/iicaiet51634.2021.9573979 · Cited by 14
Cooperative and Interactive Learning to estimate human behaviours for energy applications
da Silva, J. K. A., Amayri, M., Ploix, S., Reignier, P., & Silva, C. (2021) · Energy and Buildings · doi:10.1016/j.enbuild.2021.111727 · Cited by 18
Model Selection in Committees of Evolved Convolutional Neural Networks Using Genetic Algorithms
Baldominos, A., Sáez, Y., & Isasi, P. (2018) · Lecture notes in computer science · doi:10.1007/978-3-030-03493-1_39 · Cited by 15
Biosensor-Based Multimodal Deep Human Locomotion Decoding via Internet of Healthcare Things
Javeed, M., Abdelhaq, M., Algarni, A. M., & Jalal, A. (2023) · Micromachines · doi:10.3390/mi14122204 · Cited by 16
Neto, S. R. D. A., Santos, G. L., Rocha, É. D. S., Bendechache, M., Rosati, P., Lynn, T., & Endo, P. T. (2020) · Studies in systems, decision and control · doi:10.1007/978-3-030-38748-8_2 · Cited by 15
“The activation of a neuron is fed back to the neuron itself in order to provide a memory of past activations and to learn the temporal dynamics of time-series data [11].”
BackgroundMethodInfluentialLattanzi, E., Calisti, L., & Capellacci, P. (2023) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2023.101848 · Cited by 16
Extreme Learning Machine-Based Deep Model for Human Activity Recognition With Wearable Sensors
Niu, X., Wang, Z., & Pan, Z. (2018) · Computing in Science & Engineering · doi:10.1109/mcse.2018.110145933 · Cited by 17
Detection of human respiration patterns using deep convolution neural networks
Petrenko, A., Kyslyi, R., & Pysmennyi, I. (2018) · Eastern-European Journal of Enterprise Technologies · doi:10.15587/1729-4061.2018.139997 · Cited by 13
Sensors-Based Human Activity Recognition Using Hybrid Features and Deep Capsule Network
Ghafoor, H. Y., Jahangir, R., Jaffar, A., Alroobaea, R., Saidani, O., & Alhayan, F. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3402314 · Cited by 14
“For instance, the study [28] introduced a DeepConvLSTM comprised of CLs and LSTM layer.”
MethodArrotta, L., Bettini, C., & Civitarese, G. (2022) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-030-94822-1_25 · Cited by 34
“In particular, we slightly adapted the method presented in [16] to include both inertial and environmental sensors.”
MethodDomain Adaptation with Representation Learning and Nonlinear Relation for Time Series
Hussein, A., & Hajj, H. M. (2022) · ACM Transactions on Internet of Things · doi:10.1145/3502905 · Cited by 15
“It has been shown in the literature that using a hybrid network that combines convolution layers (CNN) followed by a recurrent layer sequentially provides efficient feature extraction for time series data and models temporal dynamics [21, 33].”
BackgroundImproving activity data collection with on-device personalization using fine-tuning
Mairittha, N., Mairittha, T., & Inoue, S. (2020) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3410530.3414370 · Cited by 13
Prototype Similarity Learning for Activity Recognition
Bai, L., Yao, L., Wang, X., Kanhere, S. S., & Xiao, Y. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-47426-3_50 · Cited by 14
“The performance of these methods normally depends on the effectiveness of the extracted features where are heuristic, task-independent, and not specially designed for HAR [12].”
MethodResultDomain generalization with semi-supervised learning for people-centric activity recognition
Liu, J., Zhu, W., Li, D., Hu, X., & Song, L. (2024) · Science China Information Sciences · doi:10.1007/s11432-022-3860-y · Cited by 18
Robust Finger Interactions with COTS Smartwatches via Unsupervised Siamese Adaptation
Chen, W., Wang, Z., Quan, P., Peng, Z., Lin, S., Srivastava, M. B., Matusik, W., & Stankovic, J. A. (2023) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3586183.3606794 · Cited by 20
“There are some works using smartwatches for human activity [6, 8], hand location [27, 49], gesture [5, 40, 42, 43, 53, 55, 76] or finger [45] classification.”
BackgroundSampling rate dependency in pedestrian walking speed estimation using DualCNN-LSTM
Yoshida, T., Nozaki, J., Urano, K., Hiroi, K., Kaji, K., Yonezawa, T., & Kawaguchi, N. (2019) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3341162.3343765 · Cited by 13
“To model the walking speed, we employ CNN-LSTM[18] which is successfully used for activity recognition and other temporal signal processing methods.”
MethodInfluentialRen, Y., Zhu, D., Tong, K., Xv, L., Wang, Z., Kang, L., & Chai, J. (2023) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2023.101868 · Cited by 16
Scheurer, S., Tedesco, S. A., O’Flynn, B., & Brown, K. N. (2020) · Sensors · doi:10.3390/s20133647 · Cited by 21
“On the OPPORT dataset, the state of the art achieves a subject-dependent error rate of 18% [5] and F-score of 93% [6,7].”
BackgroundIrcio, J., Lojo, A., Lozano, J. A., et al. (2022) · IEEE Computational Intelligence Magazine · doi:10.1109/mci.2021.3129962 · Cited by 13
“Therefore, for building the network architecture, it is necessary to use specific layers for MTS [29]–[32] instead of the original layers that, in [28], were specifically for images.”
MethodUnsupervised online change point detection in high-dimensional time series
Zameni, M., Sadri, A., Ghafoori, Z., Moshtaghi, M., Salim, F. D., Leckie, C. A., & Ramamohanarao, K. (2019) · Knowledge and Information Systems · doi:10.1007/s10115-019-01366-x · Cited by 15
Nisar, M. A., Shirahama, K., Irshad, M. T., Huang, X., & Grzegorzek, M. (2023) · Sensors · doi:10.3390/s23198234 · Cited by 17
“In the study presented by Morales et al. [31] related to multimodal human activity detection based on wearable sensors, they used deep convolutional and LSTM networks.”
BackgroundTuretta, C., Demrozi, F., & Pravadelli, G. (2024) · IEEE Access · doi:10.1109/access.2024.3496497 · Cited by 27
Nawaratne, R., Alahakoon, D., Silva, D. D., O’Halloran, P., Montoye, A. H. K., Staley, K., Nicholson, M., & Kingsley, M. (2020) · Journal of Sports Sciences · doi:10.1080/02640414.2020.1841394 · Cited by 13
“(Kingsley et al., 2019; Ordóñez & Roggen, 2016) The CNN model produced total weekly SB+LPA and MVPA not significantly different than the hip-specific reference method.”
MethodBenchmarking Deep Learning for Time Series: Challenges and Directions
Huang, X., Fox, G., Serebryakov, S., Mohan, A., Morkisz, P. M., & Dutta, D. (2019) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata47090.2019.9005496 · Cited by 17
“Health Medical sensors: EEG, ECG, ERP, Patient data analysis CNN, RNN 1, Markov OPPORTUNITY[17], [18], EEG[19]–[22], MIMIC[23] [24]–[30]”
MethodHuman Activity Recognition: A Dynamic Inductive Bias Selection Perspective
Hamidi, M., & Osmani, A. (2021) · Sensors · doi:10.3390/s21217278 · Cited by 17
Munir, M., Siddiqui, S. A., Küsters, F., Mercier, D., Dengel, A., & Ahmed, S. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-30493-5_43 · Cited by 32
Zou, H., Chen, Z., Zhang, C., Yuan, A., Wang, B., Wang, L., Li, J., & Pan, Y. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3373444 · Cited by 15
“By leveraging the strengths of different methods and overcoming the limitations of the traditional end-to-end models, hybrid WHAR models can further adapt to different scenarios and requirements [14].”
MethodTong, Y., & Yu, B. (2022) · Entropy · doi:10.3390/e24060845 · Cited by 15
Facilitating Human Activity Data Annotation via Context-Aware Change Detection on Smartwatches
Akbari, A., Martinez, J., & Jafari, R. (2021) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3431503 · Cited by 13
“Abundant precise annotations for this scenario become more critical when working with data hungry deep learning models [7, 8].”
BackgroundCheng, W.-Y., Lipsmeier, F., Scotland, A., Creagh, A., Kilchenmann, T., Jin, L., Schjodt-Eriksen, J., Wolf, D., Zhang-Schaerer, Y.-P., Garcia, I. F., et al. (2017) · 2017 IEEE Life Sciences Conference (LSC) · doi:10.1109/lsc.2017.8268169 · Cited by 13
“The HAR model is composed of a 9-layer convolutional recurrent neural network (CRNN) adapted from previously published structure [10] and trained”
MethodEnhancing Representation of Deep Features for Sensor-Based Activity Recognition
Li, X., Nie, L., Si, X., Ding, R., & Zhan, D. (2020) · Mobile Networks and Applications · doi:10.1007/s11036-020-01689-y · Cited by 13
“…deep learning approaches, including the widely used restricted Boltzmann machines [26], CNN [27, 28], recurrent neural networks (RNN) [29], LSTM [30] and their composition models [31, 32] have been applied and proven to be superior to the state-of-the-art methods in a number of recognition tasks.”
MethodUbi-Asthma: Toward Ubiquitous Asthma Detection Using the Smartwatch
Wu, Y. L., Zhang, J., Chen, Y., Wang, J., Shi, W., & Zhang, Q. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3243188 · Cited by 16
“We adopt a stateof-the-art convolutional network [40] as the feature extractor to compute the backbone features.”
MethodIdentification of Walker Identity Using Smartphone Sensors: An Experiment Using Ensemble Learning
Angrisano, A., Bernardi, M. L., Cimitile, M., Gaglione, S., & Vultaggio, M. (2020) · IEEE Access · doi:10.1109/access.2020.2971693 · Cited by 13
Li, Y., Wang, L., & Liu, F. (2022) · Electronics · doi:10.3390/electronics11162526 · Cited by 14
Advancing Activity Recognition With Multimodal Fusion and Transformer Techniques
Aidarova, S., Nurmakhan, T., Myrzakhan, R., Fazli, S., & Yazıcı, A. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3555663 · Cited by 18
“Ordóñez and Roggen [34] also developed a CNN–RNN model named DeepConvLSTM.”
Rodríguez, M. H., Ruiz, L. G. B., Criado-Ramón, D., & Pegalajar, M. C. (2023) · Machine Learning and Knowledge Extraction · doi:10.3390/make5020026 · Cited by 18
“LSTM networks are a type of recurrent neural network (RNN) designed to handle the issue of vanishing gradients in traditional RNNs. LSTMs are well suited for tasks involving sequences of data, such as time series prediction, language translation, and speech recognition [24,30,45].”
Chen, X., Zhou, X., Sun, M., & Wang, H. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3491933 · Cited by 13
“Supervised learning models for HAR have been widely developed and achieve promising performance [1], [3], [20], [21].”
A systematic review of abnormal behaviour detection and analysis in driving simulators
Tang, Y. M., Zhao, D., Chen, T., et al. (2025) · Transportation Research Part F Traffic Psychology and Behaviour · doi:10.1016/j.trf.2025.01.002 · Cited by 13
Paszek, K., & Grzechca, D. E. (2023) · Sensors · doi:10.3390/s23198270 · Cited by 15
An Overview of Machine Learning and 5G for People with Disabilities
Domingo, M. C. (2021) · Sensors · doi:10.3390/s21227572 · Cited by 18
Smart Suspenders With Sensors and Machine Learning for Human Activity Monitoring
Mani, N., Haridoss, P., & George, B. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3263231 · Cited by 13
“Other widely used models, such as the random forest (RF) [30], [31], decision tree (DT) [32], [33], logistic regression (LR) [34], [35], and neural networks, including long short-term memory (LSTM) models [36], [37], [38], [39], [40], [41], [42], are part of standard HAR classifier.”
BackgroundMethodSubset Replay Based Continual Learning for Scalable Improvement of Autonomous Systems
Brahma, P. P., & Othon, A. (2018) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2018.00154 · Cited by 16
“Deep learning has stood out to be the stateof-the-art for object classification [15], activity recognition [3] [23], semantic segmentation [25] [20] and depth estimation [18] [4].”
BackgroundLiu, X. (2022) · Computer Communications · doi:10.1016/j.comcom.2022.07.006 · Cited by 14
Recurrent neural system with minimum complexity: A deep learning perspective
Sun, X., Li, T., Li, Y., Li, Q., Huang, Y., & Liu, J. (2017) · Neurocomputing · doi:10.1016/j.neucom.2017.09.075 · Cited by 14
A Deep Learning based Intrusion Detection Approach for MANET
(2022) · doi:10.21203/rs.3.rs-1349334/v1 · Cited by 63
Alshammari, S. A., & Albalawi, N. S. (2024) · Engineering Technology & Applied Science Research · doi:10.48084/etasr.9255 · Cited by 15
Radar for assisted living in the context of Internet of Things for Health and beyond
Kernec, J. L., Fioranelli, F., Yang, S., Lorandel, J., & Romain, O. (2018) · IFIP/IEEE International Conference on Very Large Scale Integration · doi:10.1109/vlsi-soc.2018.8644816 · Cited by 13
“[51] shows a combination of CNN and LSTM for wearables and [49] presents a multimodal CNN multi-stream in parallel with LSTM with fusion; showing new ways to think about classifying data as a continuum.”
BackgroundIdentify Significant Phenomenon-Specific Variables for Multivariate Time Series
Hao, Y., Cao, H., Mueen, A., & Brahma, S. (2019) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2019.2934464 · Cited by 12
“Second, CNN has shown good performance in classifying large amount of data in very high dimensional space [16], [17]; thus adopting CNN can help reduce the computational complexity.”
MethodDas, S., Meher, S., & Sahoo, U. K. (2022) · Sensors · doi:10.3390/s22113968 · Cited by 16
Annotation Performance for multi-channel time series HAR Dataset in Logistics
Reining, C., Rueda, F. M., Niemann, F., Fink, G. A., & ten Hompel, M. (2020) · IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops) · doi:10.1109/percomworkshops48775.2020.9156170 · Cited by 15
“Due to the intra- and inter-class variability of human motion, a large amount of observations from different subjects is necessary for methods of supervised learning [2].”
BackgroundDiscrimination of cycling patterns using accelerometric data and deep learning techniques
Procházka, A., Charvátová, H., Vyšata, O., Jarchi, D., & Sanei, S. (2020) · Neural Computing and Applications · doi:10.1007/s00521-020-05504-3 · Cited by 14
“The classification system included the input layer, bidirectional long short-term memory (LSTM) [43], fully connected layer, softmax layer, and the classification layer [19, 41], as presented in Fig.”
MethodHuman behavior recognition based on sparse transformer with channel attention mechanism
Cao, K., & Wang, M. (2023) · Frontiers in Physiology · doi:10.3389/fphys.2023.1239453 · Cited by 12
Yoshimura, N., Maekawa, T., & Hara, T. (2021) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn52387.2021.9533888 · Cited by 12
“Based on the results of neural network-based activity recognition [1], [26], the networks appeared to be properly trained.”
BackgroundFluid intake recognition using inertial sensors
Wellnitz, A., Wolff, J.-P., Haubelt, C. D., & Kirste, T. (2019) · International Workshop on Sensor-based Activity Recognition and Interaction (iWOAR) · doi:10.1145/3361684.3361688 · Cited by 15
“The combination of convolutional and recurrent neural networks has been successfully shown to result in a very high-performing architecture for the recognition of several gestures (including drinking) and activities[10].”
BackgroundComplex Human Activities Recognition Based on High Performance 1D CNN Model
Maurya, R., Teo, T. H., Chua, S. H., Chow, H., & Wey, I. (2022) · IEEE International Symposium on Embedded Multicore/Many-core Systems-on-Chip (MCSoC) · doi:10.1109/mcsoc57363.2022.00059 · Cited by 14
Detection of Dementia-Related Abnormal Behaviour Using Recursive Auto-Encoders
Arifoğlu, D., Wang, Y., & Bouchachia, A. (2021) · Sensors · doi:10.3390/s21010260 · Cited by 11
“In [35], convolutional neural networks (CNNs) and long short term (LSTM) recurrent neural networks (RNNs) were used to recognise activities from wearable sensors.”
Method(2024) · doi:10.2196/59469 · Cited by 14
“We focused on DL architectures commonly used for wearable sensor data, such as CNN 1D [21,22], CNN 2D [21], attention model [23], and convolutional long short-term memory (ConvLSTM) [24].”
MethodTransNet: Minimally Supervised Deep Transfer Learning for Dynamic Adaptation of Wearable Systems
Rokni, S. A., Nourollahi, M., Alinia, P., Mirzadeh, I., Pedram, M., & Ghasemzadeh, H. G. (2020) · ACM Transactions on Design Automation of Electronic Systems · doi:10.1145/3414062 · Cited by 15
“Prior research suggests that deeper networks with more parameters offer a higher learning capacity and can potentially achieve a higher prediction performance in the source domain at the expense of higher computation and training time [56, 57].”
BackgroundMethodInfluentialReining, C., Rueda, F. M., ten Hompel, M., & Fink, G. A. (2018) · Annals of Computer Science and Information Systems · doi:10.15439/2018f188 · Cited by 15
“Furthermore, due to the large intra- and inter-class variability of the human movements, a large number of experiments must be carried out, which draw motion repetitions from the same or different persons [7].”
BackgroundMethodInfluentialFortes Rey, V., Suh, S., & Lukowicz, P. (2022) · Proceedings of the 2022 ACM International Symposium on Wearable Computers · doi:10.1145/3544794.3558464 · Cited by 12
“Over the past few years, deep learning methods have emerged as the state-of-the-art approaches in HAR by utilizing convolutional neural networks (CNN) and long short-term memory networks (LSTM) [1, 4, 10, 13, 16].”
BackgroundA Deep Learning Framework for Construction Equipment Activity Analysis
Hernandez, C., Slaton, T., Balali, V., et al. (2019) · ASCE International Conference on Computing in Civil Engineering (i3CE) · doi:10.1061/9780784482438.061 · Cited by 12
Super Low Resolution RF Powered Accelerometers for Alerting on Hospitalized Patient Bed Exits
Chesser, M., Jayatilaka, A., Visvanathan, R., Fumeaux, C., Sample, A. P., & Ranasinghe, D. C. (2019) · IEEE International Conference on Pervasive Computing and Communications · doi:10.1109/percom.2019.8767398 · Cited by 21
“This differs slightly from the original network described in [32].”
Tian, Y., Hettiarachchi, D., & Kamijo, S. (2022) · Sensors · doi:10.3390/s22176453 · Cited by 12
“Ord ó ñez et al. present a Deep-Conv LSTM framework [21] composed of convolutional layers and an LSTM layer, this inputs the time domain data of multiple sensors into the convolutional layer without pre-processing, and further processes the output of the convolutional layer by the LSTM layer.”
BackgroundMethodLiu, M., Geißler, D., Nshimyimana, D., Bian, S., Zhou, B., & Lukowicz, P. (2024) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3675094.3678446 · Cited by 13
“…community has come up with the idea of organizing multimodal time series as „fake images“ to take advantage of CNN’s which improved performance and became a standard technique for low-level feature extraction [10] but the improvement remains much more modest than the dramatic effect CNNs had on CV.”
BackgroundEvolutionary Design of Recurrent Neural Network Architecture for Human Activity Recognition
Viswambaran, R. A., Chen, G., Xue, B., & Nekooei, S. M. (2019) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec.2019.8790050 · Cited by 14
“7 CNNs [35] UCI-Opp 88.”
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.