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 601–700 of 2,807
Contrastive Accelerometer–Gyroscope Embedding Model for Human Activity Recognition
Koo, I., Park, Y., Jeong, M., & Kim, C. (2022) · IEEE Sensors Journal · doi:10.1109/jsen.2022.3222825 · Cited by 40
“Many deep-learning-based HAR systems using multimodal sensors take the data-level fusion method [24], [25], [27], [28], [29], [30], [31].”
BackgroundMethodInfluentialErcolano, G., & Rossi, S. (2021) · Intelligent Service Robotics · doi:10.1007/s11370-021-00358-7 · Cited by 29
“…model aims to explore the combination of CNN for representation learning and of LSTM for temporal dependencies learning, which is proposed in applications that concern spatiotemporal classification, like in [6] for video description and in [20] for activity recognition from wearable devices data.”
BackgroundMethodDeep Learning for Accelerometric Data Assessment and Ataxic Gait Monitoring
Procházka, A., Dostál, O., Cejnar, P., Mohamed, H. I., Pavelek, Z., Vališ, M., & Vyšata, O. (2021) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2021.3051093 · Cited by 40
“In the deep learning strategy, the classification system included the input layer with R = N / 2 elements in the basic case, bidirectional long short term memory (LSTM) [47], fully connected layer, softmax layer and the classification layer [45], [48] as presented in Fig.”
MethodAn intelligent quantitative trading system based on intuitionistic-GRU fuzzy neural networks
Wang, Y., & Luo, C. (2021) · Applied Soft Computing · doi:10.1016/j.asoc.2021.107471 · Cited by 35
Miao, S., Chen, L., Hu, R., & Luo, Y. (2022) · Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · doi:10.1145/3550331 · Cited by 36
“To further investigate the efficiency of the proposed framework, we implement DynamicWHAR and the SOTA models on two embedded devices, i.e. , Raspberry Pi 4 Model B and OnePlus 8.”
BackgroundMethodInfluentialBootstrapping Human Activity Recognition Systems for Smart Homes from Scratch
Hiremath, S. K., Nishimura, Y., Chernova, S., & Plötz, T. (2022) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3550294 · Cited by 48
Food Intake Detection from Inertial Sensors Using LSTM Networks
Kyritsis, K., Diou, C., & Delopoulos, A. N. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-70742-6_39 · Cited by 34
“We used binary cross-entropy loss with the RMSprop optimizer (with 10−3 learning rate) that has demonstrated high effectiveness in a recurrent network topology [7].”
MethodSmartphone Location Recognition: A Deep Learning-Based Approach
Klein, I. (2019) · Sensors · doi:10.3390/s20010214 · Cited by 39
A Real-Time Back-Analysis Technique to Infer Rheological Parameters from Field Monitoring
Qi, C., & Fourie, A. B. (2018) · Rock Mechanics and Rock Engineering · doi:10.1007/s00603-018-1513-2 · Cited by 35
User Authentication via Electrical Muscle Stimulation
Chen, Y., Yang, Z., Abbou, R., Lopes, P., Zhao, B. Y., & Zheng, H.-T. (2021) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445441 · Cited by 42
“Our implementation uses a Convolutional Neural Network (CNN) for this classifcation task [58].”
MethodMachine Learning-Based Pre-Impact Fall Detection Model to Discriminate Various Types of Fall
Kim, T. H., Choi, A., Heo, H. M., Kim, K.-R., Lee, K., & Mun, J. H. (2019) · Journal of Biomechanical Engineering · doi:10.1115/1.4043449 · Cited by 35
Comparing Sampling Strategies for Tackling Imbalanced Data in Human Activity Recognition
Alharbi, F., Ouarbya, L., & Ward, J. A. (2022) · Sensors · doi:10.3390/s22041373 · Cited by 47
Baldominos, A., Sáez, Y., & Isasi, P. (2018) · Sensors · doi:10.3390/s18041288 · Cited by 34
“The maximum value reported an F1 score of 0.9185, which is significantly better than the score reported by Ordoñez and Roggen [4] (0.915, see Table 3), but still worse than the best result of the state-of-the-art [5] (0.927).”
BackgroundMethodResultInfluentialFalls Risk Classification of Older Adults Using Deep Neural Networks and Transfer Learning
Martinez, M., & Leon, P. L. D. (2019) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2019.2906499 · Cited by 41
“application of DNNs to human activity recognition has focused on the use of deep convolutional neural networks (CNNs) [20]; recurrent neural networks (RNNs) [21], [22]; and hybrid CNNRNN architectures [23].”
BackgroundBiometric Recognition Through Eye Movements Using a Recurrent Neural Network
Jia, S., Koh, D. H., Seccia, A., Antonenko, P., Lamb, R. L., Keil, A., Schneps, M. H., & Pomplun, M. (2018) · IEEE International Conference on Big Knowledge (ICBK) · doi:10.1109/icbk.2018.00016 · Cited by 45
“LSTM networks have subsequently proved to be more effective than conventional RNNs and been widely used in machine translation [17], speech recognition [18], sequence prediction [13], [19]–[21], and so on.”
BackgroundMethodDeep Convolutional Bidirectional LSTM Based Transportation Mode Recognition
Jeyakumar, J. V., Lee, E. S., Xia, Z., Sandha, S. S., Tausik, N., & Srivastava, M. B. (2018) · ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp) · doi:10.1145/3267305.3267529 · Cited by 41
“Hence, we use deep learning models to make our activity predictions as these models do not require the extraction of features [21] from the sensor values and also generalize better when compared to the classical machine learning approaches [19, 17].”
MethodInfluentialMargin-Based Deep Learning Networks for Human Activity Recognition
Lv, T., Wang, X., Jin, L., Xiao, Y., & Song, M. N. (2020) · Sensors · doi:10.3390/s20071871 · Cited by 37
Gradient Boosted Neural Decision Forest
Dong, M., Yao, L., Wang, X., Benatallah, B., Zhang, S., & Sheng, Q. Z. (2021) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2021.3133673 · Cited by 46
“They have made significant advances in dealing with image recognition [27], activity recognition [28], document classification [29], etc.”
BackgroundZhao, Y., Shao, J., Lin, X., Sun, T., Li, J., Lian, C., Lyu, X., Si, B., & Zhan, Z. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.125912 · Cited by 40
Ciliberto, M., Fortes Rey, V., Calatroni, A., Lukowicz, P., & Roggen, D. (2021) · Frontiers in Computer Science · doi:10.3389/fcomp.2021.792065 · Cited by 35
“For instance, it was used in one of the seminal work on deep learning for human activity recognition from wearable sensors Ordóñez and Roggen (2016); Morales and Roggen (2016), and it has been used recently in fi elds as varied as machine learning model compression for embedded systems Thakker et…”
MethodSelf-citationFall Detection from Electrocardiogram (ECG) Signals and Classification by Deep Transfer Learning
Butt, F. S., Blunda, L. L., Wagner, M., Schäfer, J., Medina‐Bulo, I., & Gómez‐Ullate, D. (2021) · Information · doi:10.3390/info12020063 · Cited by 40
“The benchmark HAR dataset Opportunity [37] used in [38] consists of readings from 4 participants.”
Rawat, A., Kumar, A., Upadhyay, P., & Kumar, S. (2021) · Ecological Informatics · doi:10.1016/j.ecoinf.2021.101214 · Cited by 34
Jacobsen, E. L., Teizer, J., & Wandahl, S. (2023) · Automation in Construction · doi:10.1016/j.autcon.2023.104932 · Cited by 34
Smartphone Sensor-Based Human Locomotion Surveillance System Using Multilayer Perceptron
Azmat, U., Ghadi, Y. Y., Shloul, T. A., Alsuhibany, S. A., Jalal, A., & Park, J. (2022) · Applied Sciences · doi:10.3390/app12052550 · Cited by 39
BiLSTM-based Individual Cattle Identification for Automated Precision Livestock Farming
Qiao, Y., Su, D., Kong, H., Sukkarieh, S., Lomax, S., & Clark, C. (2020) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/case48305.2020.9217026 · Cited by 47
“These cells usually have few linear interactions making the information maintaining process easier [27].”
BackgroundData Analytics in the Internet of Things: A Survey
Saleem, T. J., & Chishti, M. A. (2019) · Scalable Computing Practice and Experience · doi:10.12694/scpe.v20i4.1562 · Cited by 40
Recognizing human violent action using drone surveillance within real-time proximity
Srivastava, A., Badal, T., Garg, A., Vidyarthi, A., & Singh, R. (2021) · Journal of Real-Time Image Processing · doi:10.1007/s11554-021-01171-2 · Cited by 35
Active Deep Learning for Activity Recognition with Context Aware Annotator Selection
Hossain, H. M. S., & Roy, N. (2019) · Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3292500.3330688 · Cited by 34
Altuve, M., Lizarazo, P., & Villamizar, J. (2020) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2020.04.007 · Cited by 34
A Comparison of Machine Learning Algorithms for Fall Detection using Wearable Sensors
Zurbuchen, N., Bruegger, P., & Wilde, A. (2020) · International Conference on Artificial Intelligence in Information and Communication (ICAIIC) · doi:10.1109/icaiic48513.2020.9065205 · Cited by 38
“There is a study [23] using Recurrent Neural Networks but there are other algorithms available such as Convolutional Neural Networks with the advantage of automatic feature extraction from time series [24].”
MethodDatabase quality assessment for interactive learning: Application to occupancy estimation
Amayri, M., Ploix, S., Bouguila, N., & Würtz, F. (2019) · Energy and Buildings · doi:10.1016/j.enbuild.2019.109578 · Cited by 34
Ji, A., Wang, Y., Miao, X., Fan, T., Ru, B., Liu, L., Nie, R., & Qiu, S. (2023) · Sensors · doi:10.3390/s23156693 · Cited by 53
Wu, Y., Liu, M., Peng, Z., Liu, M., Wang, M., & Peng, Y. (2022) · Agriculture · doi:10.3390/agriculture12081237 · Cited by 33
Learning Attribute Representation for Human Activity Recognition
Rueda, F. M., & Fink, G. A. (2018) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr.2018.8545146 · Cited by 41
“Following [3], [20], sessions ADL3 of participants 2”
BackgroundMethodInfluentialMohmed, G., Lotfi, A., & Pourabdollah, A. (2020) · International Conference on PErvasive Technologies Related to Assistive Environments (PETRA) · doi:10.1145/3389189.3397991 · Cited by 37
“The works in [20] and [3] propose an activity recognition clas-si � er based on DCNN and Long Short-term Memory (LSTM).”
MethodKhan, Y. A., Imaduddin, S., Singh, Y. P., Wajid, M., Usman, M. J., & Abbas, M. A. E. (2023) · Sensors · doi:10.3390/s23031275 · Cited by 36
Empirical comparison of deep learning models for fNIRS pain decoding
Rojas, R. F., Joseph, C., Bargshady, G., & Ou, K. (2024) · Frontiers in Neuroinformatics · doi:10.3389/fninf.2024.1320189 · Cited by 44
“This is done with the idea that differences in performance between the individual and combined models result from architectural differences rather than better pre-processing or ad hoc customisations (Ordóñez and Roggen, 2016).”
BackgroundLatent feature learning for activity recognition using simple sensors in smart homes
Chen, G., Wang, A., Zhao, S., Liu, L., & Chang, C. (2017) · Multimedia Tools and Applications · doi:10.1007/s11042-017-5100-4 · Cited by 35
“To better capture the temporal dynamics of human activities, Ordóñez and Roggen proposed to construct convolutional LSTM recurrent neural networks for activity recognition [17], which is suitable for both homogeneous and heterogeneous wearable sensor data analyses.”
MethodHuang, P., Zhang, X., Yu, S., & Guo, L. (2021) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2021.3110480 · Cited by 43
“and further used for classifying unknown ones [19], [20], [21], [22], [23]; 2) Model-based Recognition.”
A Novel Multichannel Dilated Convolution Neural Network for Human Activity Recognition
Lin, Y., & Wu, J. (2020) · Mathematical Problems in Engineering · doi:10.1155/2020/5426532 · Cited by 34
“In the traditional CNN, the pooling operation can make the convolution kernel get a larger receptive field, but it is not a strict component of CNN actually [40].”
BackgroundMethodComes, M. C., Fucci, L., Mele, F., Bove, S., Cristofaro, C., De Risi, I., Fanizzi, A., Milella, M., Strippoli, S., Zito, A., et al. (2022) · Scientific Reports · doi:10.1038/s41598-022-24315-1 · Cited by 33
Coelho, R. M., Gouveia, J. B., Botto, M. A., Krebs, H. I., & Martins, J. M. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.117306 · Cited by 39
Pham, C., Nguyen, L., Nguyen, A., Nguyen, C.-N., & Nguyen, V.-T. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-11058-w · Cited by 38
Awais, M., Chiari, L., Ihlen, E. A. F., Helbostad, J. L., & Palmerini, L. (2021) · Sensors · doi:10.3390/s21144669 · Cited by 34
“However, these deep learning-based systems were mostly trained and tested on young adults [22,26,29,30], while very few systems have been developed for older adults focusing on PAC [31] and falls [32].”
BackgroundProgressive Cross-modal Knowledge Distillation for Human Action Recognition
Ni, J., Ngu, A. H. H., & Yan, Y. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548238 · Cited by 49
“RNN type of model was also suggested to deal with time-dependent input sequences [44, 45, 54, 89].”
BackgroundInferring physical agitation in dementia using smartwatch and sequential behavior models
Alam, R., Anderson, M. S., Bankole, A., & Lach, J. (2018) · IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI ...) · doi:10.1109/bhi.2018.8333396 · Cited by 38
“For example, fall detection in elderly population [5, 15], dementia patients wandering risk monitoring [16], and for recognizing gestures and activities of daily living (ADL) [8, 10].”
BackgroundInfluentialDeep Temporal Conv-LSTM for Activity Recognition
Noor, M. H. M., Tan, S. Y., & Wahab, M. N. A. (2022) · Neural Processing Letters · doi:10.1007/s11063-022-10799-5 · Cited by 44
“Ordóñez & Roggen first proposed a novel DNN framework for activity recognition, consisting of four convolutional layers, followed by two recurrent layers and a softmax layer as a classifier [20].”
BackgroundEnhancing Deep Learning with Semantics: an application to manufacturing time series analysis
Huang, X., Zanni-Merk, C., & Crémilleux, B. (2019) · Procedia Computer Science · doi:10.1016/j.procs.2019.09.198 · Cited by 38
Otsuka, R., Yoshimura, N., Tanigaki, K., Koyama, S., Mizutani, Y., Yoda, K., & Maekawa, T. (2024) · Methods in Ecology and Evolution · doi:10.1111/2041-210x.14294 · Cited by 39
“• DCL: The original DCL has two LSTM layers after four convolution layers (Ordóñez & Roggen, 2016 ), but our DCL has one LSTM layer, following Singh et al. ( 2021 ) and Yoshimura, Morales, et al. ( 2022 ).”
BackgroundMethodA Close Look into Human Activity Recognition Models using Deep Learning
Tee, W. Z., Dave, R., Seliya, J., & Vanamala, M. (2022) · International Conference on Computing, Networks and Internet of Things (CNIOT) · doi:10.1109/cniot55862.2022.00043 · Cited by 42
Tsiktsiris, D., Dimitriou, N., Lalas, A. X., Dasygenis, M., Votis, K., & Tzovaras, D. K. (2020) · Sensors · doi:10.3390/s20174943 · Cited by 35
“Ordóñez et al. [15] have also proposed techniques for wearable human activity recognition using multimodal deep convolutional and LSTM recurrent neural networks.”
BackgroundGait identification and authentication using LSTM based on 3-axis accelerations of smartphone
Watanabe, Y., & Kimura, M. (2020) · Procedia Computer Science · doi:10.1016/j.procs.2020.09.001 · Cited by 36
Khaertdinov, B., Asteriadis, S., & Ghaleb, E. (2022) · IEEE Transactions on Biometrics Behavior and Identity Science · doi:10.1109/tbiom.2022.3180591 · Cited by 40
“They include but are not limited to Convolutional Neural Networks (CNNs) [27], Recurrent Neural Networks (RNNs) [28], [29] and their combinations [30].”
BackgroundLv, T., Wang, X., Jin, L., Xiao, Y., & Song, M. N. (2020) · IEEE Access · doi:10.1109/access.2020.2986246 · Cited by 33
“Ordóñez and Roggen showed that a hybrid architecture based on both convolutional and LSTM recurrent units functions better than do deep non-recurrent networks, and confirmed the improved performance on two benchmark datasets [6].”
BackgroundMethodInfluentialImproving Sleep Disorder Diagnosis Through Optimized Machine Learning Approaches
Rahman, M. A., Jahan, I., Islam, M., Jabid, T., Ali, M. S., Rashid, M. R. A., Islam, M. M., Ferdaus, M. H., Rasel, M. K., Jahan, M. R., et al. (2025) · IEEE Access · doi:10.1109/access.2025.3535535 · Cited by 42
Fortune, E., Cloud, B. A., Madansingh, S. I., et al. (2019) · Journal of Electromyography and Kinesiology · doi:10.1016/j.jelekin.2019.07.007 · Cited by 32
Federated Multi-Task Attention for Cross-Individual Human Activity Recognition
Shen, Q., Feng, H., Song, R., Teso, S., Giunchiglia, F., & Xu, H. (2022) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2022/475 · Cited by 40
“Building on this insight, models like DeepConvLSTM [Francisco and Daniel, 2016] and DeepSense [Yao et al.”
Recognition of Human Hand Activities Based on a Single Wrist IMU Using Recurrent Neural Networks
Rivera, P., Añazco, E. V., Choi, M., & Kim, T. (2017) · International Journal of Pharma Medicine and Biological Sciences · doi:10.18178/ijpmbs.6.4.114-118 · Cited by 37
“Recently sequential modeling approach has been employed with favorable results via recurrent neural networks (RNNs) [4], [8], [9].”
MethodCreagh, A. P., Lipsmeier, F., Lindemann, M., & Vos, M. D. (2021) · Scientific Reports · doi:10.1038/s41598-021-92776-x · Cited by 35
Experience with an Affective Robot Assistant for Children with Hearing Disabilities
Uluer, P., Köse, H., Gümüşlü, E., & Barkana, D. E. (2021) · International Journal of Social Robotics · doi:10.1007/s12369-021-00830-5 · Cited by 45
Yin, F., & Shi, F. (2021) · International Journal of Parallel Programming · doi:10.1007/s10766-021-00717-y · Cited by 43
Deep learning of smartphone sensor data for personal health assistance
Li, H., & Trocan, M. (2019) · Microelectronics Journal · doi:10.1016/j.mejo.2018.01.015 · Cited by 30
Haresamudram, H., Tang, C. I., Suh, S., Lukowicz, P., & Plötz, T. (2025) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3729467 · Cited by 33
“A combination of convolutional and recurrent layers was proposed in DeepConvLSTM [137], which, even today, remains a strong baseline and is commonly utilized in contemporary HAR works [77, 181].”
Emotion Recognition using EEG and Physiological Data for Robot-Assisted Rehabilitation Systems
Gümüşlü, E., Barkana, D. E., & Köse, H. (2020) · Companion Publication of the International Conference on Multimodal Interaction (ICMI Companion) · doi:10.1145/3395035.3425199 · Cited by 40
“A Multi-Layer Convolutional Neural Network (CNN) with denoising Auto-Encoders [29], a convolutional long short-term memory network with a temporal margin-based loss function [25], CNN and stacked autoencoders (SAE) [50], multichannels deep CNN and long short-term memory recurrent neural network [16, 32], Deep Convolut…”
MethodHuman worker activity recognition in a production floor environment through deep learning
Mastakouris, A., Andriosopoulou, G., Masouros, D., Benardos, P., Vosniakos, G.-C., & Soudris, D. (2023) · Journal of Manufacturing Systems · doi:10.1016/j.jmsy.2023.08.020 · Cited by 34
BeSafe B2.0 Smart Multisensory Platform for Safety in Workplaces
Márquez-Sánchez, S., Campero-Jurado, I., Robles-Camarillo, D., Rodríguez, S., & Corchado, J. M. (2021) · Sensors · doi:10.3390/s21103372 · Cited by 40
“Therefore, Machine Learning-based models are required to learn from the data to extract knowledge and detect and classify behavioral patterns [55–57].”
BackgroundImproved LSTM-Based Abnormal Stream Data Detection and Correction System for Internet of Things
Liu, J., Bai, J., Li, H., & Sun, B. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2021.3079504 · Cited by 40
Application of deep learning to improve sleep scoring of wrist actigraphy
Haghayegh, S., Khoshnevis, S., Smolensky, M. H., & Diller, K. R. (2020) · Sleep Medicine · doi:10.1016/j.sleep.2020.05.008 · Cited by 31
Application of Deep Learning for Quality of Service Enhancement in Internet of Things: A Review
Kimbugwe, N., Pei, T., & Kyebambe, M. N. (2021) · Energies · doi:10.3390/en14196384 · Cited by 35
“RNNs are capable of taking one or more input vectors and generating more output vectors, and unlike NN, where outputs are only determined by the weights of input vectors, they also use hidden state vectors, which show the context of the previous data [15,84].”
BackgroundInertial-Measurement-Unit-Based Novel Human Activity Recognition Algorithm Using Conformer
Kim, Y.-W., Cho, W.-H., Kim, K.-S., & Lee, S. (2022) · Sensors · doi:10.3390/s22103932 · Cited by 31
“Ord ó ñez [33] proposed the DeepConvLSTM HAR model, and this model showed better performance than the CNN-based HAR model in the public HAR datasets.”
MethodA Novel Attention-Based Convolution Neural Network for Human Activity Recognition
Zheng, G. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3122258 · Cited by 33
Movement and Gesture Recognition Using Deep Learning and Wearable-sensor Technology
Xie, B., Li, B., & Harland, A. R. (2018) · doi:10.1145/3268866.3268890 · Cited by 36
Bello, H., Zhou, B., Suh, S., & Lukowicz, P. (2021) · International Symposium on Wearable Computers · doi:10.1145/3460421.3480418 · Cited by 32
“The processed data of 14 volunteers (5600 instances) were fed to several deep learning models that have previously been shown to be effective in activity/gesture recognition such as modified 1DLeNet5 [30, 46], DeepConvLSTM [36], and Conv2D [23, 24, 42].”
MethodInfluentialAdaptive Thresholding of CNN Features for Maize Leaf Disease Classification and Severity Estimation
Mafukidze, H. D., Owomugisha, G., Otim, D., Nechibvute, A., Nyamhere, C., & Mazunga, F. (2022) · Applied Sciences · doi:10.3390/app12178412 · Cited by 41
Human activity recognition adapted to the type of movement
Gil-Martín, M., San-Segundo, R., Fernández-Martínez, F., & de Córdoba, R. (2020) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2020.106822 · Cited by 36
A multi-scale feature extraction fusion model for human activity recognition
Zhang, C., Cao, K., Lu, L., & Deng, T. (2022) · Scientific Reports · doi:10.1038/s41598-022-24887-y · Cited by 46
“Ordonez and Roggen 43 combined deep CNN and LSTM for the classification of 27 gestures and 5 actions.”
BackgroundEstimating Occupancy Using Interactive Learning With a Sensor Environment: Real-Time Experiments
Amayri, M., Ploix, S., Bouguila, N., et al. (2019) · IEEE Access · doi:10.1109/access.2019.2911921 · Cited by 31
“Moreover, [15] proposed an action recognition approach based on Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs).”
MethodMeyer, J., Frank, A., Schlebusch, T., & Kasneci, E. (2021) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3494998 · Cited by 34
“However, multiple works have managed to successfully apply CNNs for HAR tasks [22, 44, 48, 61].”
BackgroundContext-aware mutual learning for semi-supervised human activity recognition using wearable sensors
Qu, Y., Tang, Y., Yang, X., Wen, Y., & Zhang, W. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.119679 · Cited by 35
A Multitask Deep Learning Approach for Sensor-Based Human Activity Recognition and Segmentation
Duan, F., Zhu, T., Wang, J., Chen, L., Ning, H., & Wan, Y. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3273673 · Cited by 51
“Ordóñez and Roggen [27] proposed combining CNN and LSTM to fuse multimodal sensor information for improving activity recognition performance.”
BackgroundMethodSteganalysis of VoIP Streams with CNN-LSTM Network
Yang, H., Yang, Z., & Huang, Y. (2019) · ACM Workshop on Information Hiding and Multimedia Security (IH&MMSec) · doi:10.1145/3335203.3335735 · Cited by 31
“Noticed that CNN is able to learn local response from temporal or spatial data but lacks the ability to learn sequential correlations, while RNN is able to handle sequences of any lengths and capture long-term contextual dependency [5, 15], in this paper, we indicate a proper way to take advantage of these two archite…”
MethodUllah, S., Pirahandeh, M., & Kim, D. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.127157 · Cited by 38
A battery management strategy in microgrid for personalized customer requirements
Chen, P., Liu, M., Chen, C., & Shang, X. C. (2019) · Energy · doi:10.1016/j.energy.2019.116245 · Cited by 29
Cheng, D., Zhang, L., Qin, L., Wang, S., Wu, H., & Song, A. (2024) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2024.3373019 · Cited by 31
“For HAR tasks, the extraction of local features is particularly crucial [17], [26], [27], [38].”
BackgroundInfluentialSmartphone Sensor-Based Human Activity Recognition Robust to Different Sampling Rates
Hasegawa, T. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.3038281 · Cited by 32
Wang, X., Zhang, L., Huang, W., Wang, S., Wu, H., He, J., & Song, A. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3132088 · Cited by 42
“In order to capture temporal dynamics for HAR, Ordóñez et al.[18] proposed a novel network architecture called DeepConvLSTM consisting of convolutional and recurrent units, which significantly outperforms CNN alone.”
BackgroundKwon, H., Abowd, G. D., & Plötz, T. (2021) · Sensors · doi:10.3390/s21248337 · Cited by 30
“Recurrent neural networks (RNNs) are often used in combination with convolutional operators in an effort to capture global temporal dynamics of the feature representations learned by the convolutional part of the model [7].”
BackgroundMethodInfluentialWalk‐to‐Charge Technology: Exploring Efficient Energy Harvesting Solutions for Smart Electronics
Beniwal, R., Kalra, S., Beniwal, N. S., Mazumdar, H., Singhal, A. K., & Singh, S. K. (2023) · Journal of Sensors · doi:10.1155/2023/6614658 · Cited by 34
“Optimization techniques include sensor selection and placement to maximize coverage and minimize redundancy, as well as data fusion algorithms tailored to the specific application to optimize information extraction and system performance [42, 43].”
MethodGrouped Convolutional Neural Networks for Multivariate Time SeriesPreprint
Yi, S., Ju, J., Yoon, M.-K., & Choi, J. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1703.09938 · Cited by 51
“Recently, variants of CNN are applied to analyzing various kinds of time-series such as sensor values and EEG (electroencephalogram) signals (Yang et al., 2015; Ordez & Roggen, 2016).”
MethodCNN-LSTM Model for Recognizing Video-Recorded Actions Performed in a Traditional Chinese Exercise
Chen, J., Wang, J., Yuan, Q., & Zhao, Y. (2023) · IEEE Journal of Translational Engineering in Health and Medicine · doi:10.1109/jtehm.2023.3282245 · Cited by 46
“[36] proposed a generic deep framework for activity recognition based on convolutional and LSTM recurrent units.”
BackgroundYu, Z., Yang, K., Luo, Y., et al. (2021) · International Journal of Applied Earth Observation and Geoinformation · doi:10.1016/j.jag.2021.102344 · Cited by 30
Learning Temporal and Bodily Attention in Protective Movement Behavior Detection
Wang, C., Peng, M., Olugbade, T., Lane, N. D., Williams, A. C. D. C., & Bianchi‐Berthouze, N. (2019) · International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW) · doi:10.1109/aciiw.2019.8925084 · Cited by 39
“[20] focused on temporal attention, which they embedded at the end of a convolutional LSTM network (Conv-LSTM) [25].”
BackgroundGaugel, S., & Reichert, M. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.106078 · Cited by 40
“[13] presented an LSTM-CNN-based network for human activity recognition based on multivariate time series data.”
BackgroundMethodPopulation and Age-Based Cardiorespiratory Fitness Level Investigation and Automatic Prediction
Xiang, L., Deng, K., Mei, Q., Gao, Z., Yang, T., Wang, A., Fernandez, J., & Gu, Y. (2022) · Frontiers in Cardiovascular Medicine · doi:10.3389/fcvm.2021.758589 · Cited by 44
Multi-Modal Deep Learning for Assessing Surgeon Technical Skill
Kasa, K., Burns, D., Goldenberg, M. G., Selim, O., Whyne, C. M., & Hardisty, M. (2022) · Sensors · doi:10.3390/s22197328 · Cited by 39
Compressing RNNs for IoT devices by 15-38x using Kronecker Products
Thakker, U., Beu, J., Gope, D., Zhou, C., Fedorov, I., Dasika, G., & Mattina, M. (2019) · arXiv · doi:10.48550/arxiv.1906.02876 · Cited by 30
“As a result, RNNs can greatly benefit important Internet-of-Things (IoT) applications like wake-word detection [54], human activity recognition [18, 38, 39] and predictive maintenance [3, 42].”
BackgroundWang, A., Zhao, S., Zheng, C., Yang, J., Chen, G., & Chang, C. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.3035062 · Cited by 44
“used deep CNNs to recognize ADLs of the elderly [30], where they took as input the activity image and applied the two-dimensional convolution to learn features.”
BackgroundMethodTransfer learning across human activities using a cascade neural network architecture
Du, X., Farrahi, K., & Niranjan, M. (2019) · International Symposium on Wearable Computers · doi:10.1145/3341163.3347730 · Cited by 33
“The most popular Deep Learning approaches applied in HAR include Multi-Layer Perceptrons (MLPs), Convolutional Neural Networks (CNNs) [23], Recurrent Neural Networks [9], and Long Short Term Memory networks (LSTMs) [11].”
MethodLogacjov, A., & Bach, K. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107478 · Cited by 41
LHAR: Lightweight Human Activity Recognition on Knowledge Distillation
Deng, S., Chen, J., Teng, D., Yang, C., Chen, D., Jia, T., & Wang, H. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3298932 · Cited by 38
“(cid:2) LSTM [37] is the teacher network based on CNN and LSTM recurrent units.”
Background
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.