{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,11]],"date-time":"2025-10-11T01:52:23Z","timestamp":1760147543557,"version":"build-2065373602"},"reference-count":45,"publisher":"MDPI AG","issue":"4","license":[{"start":{"date-parts":[[2023,2,12]],"date-time":"2023-02-12T00:00:00Z","timestamp":1676160000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>Hand gesture recognition applications based on surface electromiographic (sEMG) signals can benefit from on-device execution to achieve faster and more predictable response times and higher energy efficiency. However, deploying state-of-the-art deep learning (DL) models for this task on memory-constrained and battery-operated edge devices, such as wearables, requires a careful optimization process, both at design time, with an appropriate tuning of the DL models\u2019 architectures, and at execution time, where the execution of large and computationally complex models should be avoided unless strictly needed. In this work, we pursue both optimization targets, proposing a novel gesture recognition system that improves upon the state-of-the-art models both in terms of accuracy and efficiency. At the level of DL model architecture, we apply for the first time tiny transformer models (which we call bioformers) to sEMG-based gesture recognition. Through an extensive architecture exploration, we show that our most accurate bioformer achieves a higher classification accuracy on the popular Non-Invasive Adaptive hand Prosthetics Database 6 (Ninapro DB6) dataset compared to the state-of-the-art convolutional neural network (CNN) TEMPONet (+3.1%). When deployed on the RISC-V-based low-power system-on-chip (SoC) GAP8, bioformers that outperform TEMPONet in accuracy consume 7.8\u00d7\u201344.5\u00d7 less energy per inference. At runtime, we propose a three-level dynamic inference approach that combines a shallow classifier, i.e., a random forest (RF) implementing a simple \u201crest detector\u201d with two bioformers of different accuracy and complexity, which are sequentially applied to each new input, stopping the classification early for \u201ceasy\u201d data. With this mechanism, we obtain a flexible inference system, capable of working in many different operating points in terms of accuracy and average energy consumption. On GAP8, we obtain a further 1.03\u00d7\u20131.35\u00d7 energy reduction compared to static bioformers at iso-accuracy.<\/jats:p>","DOI":"10.3390\/s23042065","type":"journal-article","created":{"date-parts":[[2023,2,13]],"date-time":"2023-02-13T02:14:11Z","timestamp":1676254451000},"page":"2065","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Reducing the Energy Consumption of sEMG-Based Gesture Recognition at the Edge Using Transformers and Dynamic Inference"],"prefix":"10.3390","volume":"23","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9225-3106","authenticated-orcid":false,"given":"Chen","family":"Xie","sequence":"first","affiliation":[{"name":"Department of Control and Computer Engineering, Politecnico di Torino, 10129 Turin, Italy"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6215-8220","authenticated-orcid":false,"given":"Alessio","family":"Burrello","sequence":"additional","affiliation":[{"name":"Interuniversity Department of Regional and Urban Studies and Planning, Politecnico di Torino, 10129 Turin, Italy"},{"name":"Department of Electrical, Electronic and Information Engineering, University of Bologna, 40136 Bologna, Italy"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9595-7216","authenticated-orcid":false,"given":"Francesco","family":"Daghero","sequence":"additional","affiliation":[{"name":"Department of Control and Computer Engineering, Politecnico di Torino, 10129 Turin, Italy"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8068-3806","authenticated-orcid":false,"given":"Luca","family":"Benini","sequence":"additional","affiliation":[{"name":"Department of Electrical, Electronic and Information Engineering, University of Bologna, 40136 Bologna, Italy"},{"name":"Department of Information Technology and Electrical Engineering, ETH Zurich, 8092 Zurich, Switzerland"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5881-3811","authenticated-orcid":false,"given":"Andrea","family":"Calimera","sequence":"additional","affiliation":[{"name":"Department of Control and Computer Engineering, Politecnico di Torino, 10129 Turin, Italy"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9046-5618","authenticated-orcid":false,"given":"Enrico","family":"Macii","sequence":"additional","affiliation":[{"name":"Interuniversity Department of Regional and Urban Studies and Planning, Politecnico di Torino, 10129 Turin, Italy"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1369-9688","authenticated-orcid":false,"given":"Massimo","family":"Poncino","sequence":"additional","affiliation":[{"name":"Department of Control and Computer Engineering, Politecnico di Torino, 10129 Turin, Italy"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2872-7071","authenticated-orcid":false,"given":"Daniele","family":"Jahier Pagliari","sequence":"additional","affiliation":[{"name":"Department of Control and Computer Engineering, Politecnico di Torino, 10129 Turin, Italy"}]}],"member":"1968","published-online":{"date-parts":[[2023,2,12]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"68","DOI":"10.1016\/j.nurt.2007.10.069","article-title":"Responsive cortical stimulation for the treatment of epilepsy","volume":"5","author":"Sun","year":"2008","journal-title":"Neurotherapeutics"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"247","DOI":"10.1016\/bs.adcom.2020.07.002","article-title":"Energy-Efficient Deep Learning Inference on Edge Devices","volume":"Volume 122","author":"Kim","year":"2021","journal-title":"Hardware Accelerator Systems for Artificial Intelligence and Machine Learning"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"1149","DOI":"10.1109\/TCPMT.2018.2799987","article-title":"An sEMG-Based Human\u2013Robot Interface for Robotic Hands Using Machine Learning and Synergies","volume":"8","author":"Meattini","year":"2018","journal-title":"IEEE Trans. Components Packag. Manuf. Technol."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"290","DOI":"10.1109\/THMS.2022.3146787","article-title":"L-sign: Large-vocabulary sign gestures recognition system","volume":"52","author":"Zheng","year":"2022","journal-title":"IEEE Trans. Hum.-Mach. Syst."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"115657","DOI":"10.1016\/j.eswa.2021.115657","article-title":"Vision-based hand gesture recognition using deep learning for the interpretation of sign language","volume":"182","author":"Sharma","year":"2021","journal-title":"Expert Syst. Appl."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"140053","DOI":"10.1007\/s42979-021-00827-x","article-title":"Methods, databases and recent advancement of vision-based hand gesture recognition for hci systems: A review","volume":"2","author":"Sarma","year":"2021","journal-title":"SN Comput. Sci."},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Palermo, F., Cognolato, M., Gijsberts, A., Muller, H., Caputo, B., and Atzori, M. (2017, January 17\u201320). Repeatability of grasp recognition for robotic hand prosthesis control based on sEMG data. Proceedings of the 2017 International Conference on Rehabilitation Robotics (ICORR), London, UK.","DOI":"10.1109\/ICORR.2017.8009405"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"140053","DOI":"10.1038\/sdata.2014.53","article-title":"Electromyography data for non-invasive naturally-controlled robotic hand prostheses","volume":"1","author":"Atzori","year":"2014","journal-title":"Sci. Data"},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Kaufmann, P., Englehart, K., and Platzner, M. (September, January 31). Fluctuating EMG signals: Investigating long-term effects of pattern matching algorithms. Proceedings of the 2010 Annual International Conference of the IEEE Engineering in Medicine and Biology, Buenos Aires, Argentina.","DOI":"10.1109\/IEMBS.2010.5627288"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"620","DOI":"10.1109\/TBCAS.2015.2476555","article-title":"A Versatile Embedded Platform for EMG Acquisition and Gesture Recognition","volume":"9","author":"Benatti","year":"2015","journal-title":"IEEE Trans. Biomed. Circuits Syst."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Milosevic, B., Farella, E., and Benatti, S. (2018, January 26\u201329). Exploring Arm Posture and Temporal Variability in Myoelectric Hand Gesture Recognition. Proceedings of the 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob), Enschede, The Netherlands.","DOI":"10.1109\/BIOROB.2018.8487838"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Hu, Y., Wong, Y., Wei, W., Du, Y., Kankanhalli, M., and Geng, W. (2018). A novel attention-based hybrid CNN-RNN architecture for sEMG-based gesture recognition. PLoS ONE, 13.","DOI":"10.1371\/journal.pone.0206049"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Tsinganos, P., Cornelis, B., Cornelis, J., Jansen, B., and Skodras, A. (2018, January 19\u201321). Deep Learning in EMG-based Gesture Recognition. Proceedings of the 5th International Conference on Physiological Computing Systems, Seville, Spain.","DOI":"10.5220\/0006960201070114"},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Tsinganos, P., Cornelis, B., Cornelis, J., Jansen, B., and Skodras, A. (2019, January 12\u201317). Improved gesture recognition based on sEMG signals and TCN. Proceedings of the ICASSP 2019\u20142019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Brighton, UK.","DOI":"10.1109\/ICASSP.2019.8683239"},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Flamand, E., Rossi, D., Conti, F., Loi, I., Pullini, A., Rotenberg, F., and Benini, L. (2018, January 10\u201312). GAP-8: A RISC-V SoC for AI at the Edge of the IoT. Proceedings of the 2018 IEEE 29th International Conference on Application-Specific Systems, Architectures and Processors (ASAP), Milan, Italy.","DOI":"10.1109\/ASAP.2018.8445101"},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Betthauser, J.L., Krall, J.T., Kaliki, R.R., Fifer, M.S., and Thakor, N.V. (2019, January 20\u201323). Stable Electromyographic Sequence Prediction during Movement Transitions using Temporal Convolutional Networks. Proceedings of the 2019 9th International IEEE\/EMBS Conference on Neural Engineering (NER), San Francisco, CA, USA.","DOI":"10.1109\/NER.2019.8717169"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Risso, M., Burrello, A., Jahier Pagliari, D., Benatti, S., Macii, E., Benini, L., and Poncino, M. (2021, January 22\u201328). Robust and Energy-efficient PPG-based Heart-Rate Monitoring. Proceedings of the 2021 IEEE International Symposium on Circuits and Systems (ISCAS), Daegu, Republic of Korea.","DOI":"10.1109\/ISCAS51556.2021.9401282"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Jacob, B., Kligys, S., Chen, B., Zhu, M., Tang, M., Howard, A., Adam, H., and Kalenichenko, D. (2018, January 18\u201323). Quantization and training of neural networks for efficient integer-arithmetic-only inference. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00286"},{"key":"ref_19","unstructured":"Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., and Polosukhin, I. (2017). Attention is All You Need. arXiv."},{"key":"ref_20","unstructured":"Devlin, J., Chang, M.W., Lee, K., and Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv."},{"key":"ref_21","unstructured":"Brown, T.B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., and Askell, A. (2020). Language models are few-shot learners. arXiv."},{"key":"ref_22","unstructured":"Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., and Gelly, S. (2020). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv."},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Burrello, A., Morghet, F.B., Scherer, M., Benatti, S., Benini, L., Macii, E., Poncino, M., and Jahier Pagliari, D. (2022, January 14\u201323). Bioformers: Embedding transformers for ultra-low power sEMG-based gesture recognition. Proceedings of the 2022 Design, Automation & Test in Europe Conference & Exhibition (DATE), Antwerp, Belgium.","DOI":"10.23919\/DATE54114.2022.9774639"},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"244","DOI":"10.1109\/TBCAS.2019.2959160","article-title":"Robust real-time embedded emg recognition framework using temporal convolutional networks on a multicore iot processor","volume":"14","author":"Zanghieri","year":"2019","journal-title":"IEEE Trans. Biomed. Circuits Syst."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"2964","DOI":"10.1109\/TBME.2019.2899222","article-title":"Surface-electromyography-based gesture recognition by multi-view deep learning","volume":"66","author":"Wei","year":"2019","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"447","DOI":"10.1109\/TAI.2021.3098253","article-title":"A Transfer Learning Model for Gesture Recognition Based on the Deep Features Extracted by CNN","volume":"2","author":"Zou","year":"2021","journal-title":"IEEE Trans. Artif. Intell."},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Han, L., Zou, Y., and Cheng, L. (2021, January 27\u201331). A Convolutional Neural Network With Multi-scale Kernel and Feature Fusion for sEMG-based Gesture Recognition. Proceedings of the 2021 IEEE International Conference on Robotics and Biomimetics (ROBIO), Sanya, China.","DOI":"10.1109\/ROBIO54168.2021.9739426"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"82","DOI":"10.1109\/10.204774","article-title":"A new strategy for multifunction myoelectric control","volume":"40","author":"Hudgins","year":"1993","journal-title":"IEEE Trans. Bio-Med. Eng."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"848","DOI":"10.1109\/TBME.2003.813539","article-title":"A robust, real-time control scheme for multifunction myoelectric control","volume":"50","author":"Englehart","year":"2003","journal-title":"IEEE Trans. Biomed. Eng."},{"key":"ref_30","first-page":"255","article-title":"Fine detection of grasp force and posture by amputees via surface electromyography","volume":"103","author":"Castellini","year":"2009","journal-title":"J. Physiol."},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Phinyomark, A., and Scheme, E.J. (2018). EMG Pattern Recognition in the Era of Big Data and Deep Learning. Big Data Cogn. Comput., 2.","DOI":"10.3390\/bdcc2030021"},{"key":"ref_32","unstructured":"Benatti, S., Farella, E., Gruppioni, E., and Benini, L. (2014, January 3\u20136). Analysis of Robust Implementation of an EMG Pattern Recognition based Control. Proceedings of the International Joint Conference on Biomedical Engineering Systems and Technologies\u2014Volume 4. SCITEPRESS-Science and Technology Publications, Lda, Loire Valley, France."},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Cene, V.H., Tosin, M., Machado, J., and Balbinot, A. (2019). Open Database for Accurate Upper-Limb Intent Detection Using Electromyography and Reliable Extreme Learning Machines. Sensors, 19.","DOI":"10.3390\/s19081864"},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Park, E., Kim, D., Kim, S., Kim, Y.D., Kim, G., Yoon, S., and Yoo, S. (2015, January 4\u20139). Big\/Little Deep Neural Network for Ultra Low Power Inference. Proceedings of the 2015 International Conference on Hardware\/Software Codesign and System Synthesis (CODES + ISSS), Amsterdam, The Netherlands.","DOI":"10.1109\/CODESISSS.2015.7331375"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Tann, H., Hashemi, S., Bahar, R.I., and Reda, S. (2016, January 2\u20137). Runtime Configurable Deep Neural Networks for Energy-Accuracy Trade-Off. Proceedings of the Eleventh IEEE\/ACM\/IFIP International Conference on Hardware\/Software Codesign and System Synthesis\u2014CODES\u201916, Pittsburgh, PA, USA.","DOI":"10.1145\/2968456.2968458"},{"key":"ref_36","unstructured":"Yu, J., Yang, L., Xu, N., Yang, J., and Huang, T. (2018). Slimmable Neural Networks. arXiv."},{"key":"ref_37","doi-asserted-by":"crossref","unstructured":"Jahier Pagliari, D., Macii, E., and Poncino, M. (2018, January 23\u201325). Dynamic Bit-width Reconfiguration for Energy-Efficient Deep Learning Hardware. Proceedings of the International Symposium on Low Power Electronics and Design, Seattle, WA, USA.","DOI":"10.1145\/3218603.3218611"},{"key":"ref_38","doi-asserted-by":"crossref","unstructured":"Parsa, M., Panda, P., Sen, S., and Roy, K. (2017, January 11\u201315). Staged Inference Using Conditional Deep Learning for Energy Efficient Real-Time Smart Diagnosis. Proceedings of the 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Jeju, Republic of Korea.","DOI":"10.1109\/EMBC.2017.8036767"},{"key":"ref_39","doi-asserted-by":"crossref","unstructured":"Daghero, F., Jahier Pagliari, D., and Poncino, M. (2022, January 12\u201315). Two-stage Human Activity Recognition on Microcontrollers with Decision Trees and CNNs. Proceedings of the 2022 17th Conference on Ph.D Research in Microelectronics and Electronics (PRIME), Villasimius, Italy.","DOI":"10.1109\/PRIME55000.2022.9816745"},{"key":"ref_40","doi-asserted-by":"crossref","unstructured":"Xie, C., Jahier Pagliari, D., and Calimera, A. (2022, January 12\u201315). Energy-efficient and Privacy-aware Social Distance Monitoring with Low-resolution Infrared Sensors and Adaptive Inference. Proceedings of the 2022 17th Conference on Ph.D Research in Microelectronics and Electronics (PRIME), Villasimius, Italy.","DOI":"10.1109\/PRIME55000.2022.9816801"},{"key":"ref_41","unstructured":"Mullapudi, R.T., Mark, W.R., Shazeer, N., and Fatahalian, K. (2018, January 18\u201323). HydraNets: Specialized Dynamic Architectures for Efficient Inference. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, UT, USA."},{"key":"ref_42","doi-asserted-by":"crossref","first-page":"19","DOI":"10.1145\/3487910","article-title":"Embedding Temporal Convolutional Networks for Energy-Efficient PPG-Based Heart Rate Monitoring","volume":"3","author":"Burrello","year":"2022","journal-title":"ACM Trans. Comput. Healthc."},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Daghero, F., Burrello, A., Jahier Pagliari, D., Benini, L., Macii, E., and Poncino, M. (2020, January 23\u201325). Energy-Efficient Adaptive Machine Learning on IoT End-Nodes With Class-Dependent Confidence. Proceedings of the 2020 27th IEEE International Conference on Electronics, Circuits and Systems (ICECS), Glasgow, UK.","DOI":"10.1109\/ICECS49266.2020.9294863"},{"key":"ref_44","doi-asserted-by":"crossref","unstructured":"Burrello, A., Scherer, M., Zanghieri, M., Conti, F., and Benini, L. (2021, January 23\u201325). A Microcontroller is All You Need: Enabling Transformer Execution on Low-Power IoT Endnodes. Proceedings of the 2021 IEEE International Conference on Omni-Layer Intelligent Systems (COINS), Barcelona, Spain.","DOI":"10.1109\/COINS51742.2021.9524173"},{"key":"ref_45","doi-asserted-by":"crossref","unstructured":"Garofalo, A., Rusci, M., Conti, F., Rossi, D., and Benini, L. (2019). PULP-NN: Accelerating Quantized Neural Networks on Parallel Ultra-Low-Power RISC-V Processors. arXiv.","DOI":"10.23919\/DATE48585.2020.9116529"}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/23\/4\/2065\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T18:32:07Z","timestamp":1760121127000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/23\/4\/2065"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,2,12]]},"references-count":45,"journal-issue":{"issue":"4","published-online":{"date-parts":[[2023,2]]}},"alternative-id":["s23042065"],"URL":"https:\/\/doi.org\/10.3390\/s23042065","relation":{},"ISSN":["1424-8220"],"issn-type":[{"type":"electronic","value":"1424-8220"}],"subject":[],"published":{"date-parts":[[2023,2,12]]}}}