XGest: Enabling Cross-Label Gesture Recognition with RF Signals
The DOI matches its Crossref record; 39 of the 39 references it cites were found in a registry or an index.
Abstract
Extensive efforts have been devoted to human gesture recognition with radio frequency (RF) signals. However, their performance degrades when applied to novel gesture classes that have never been seen in the training set. To handle unseen gestures, extra efforts are inevitable in terms of data collection and model retraining. In this article, we present XGest, a cross-label gesture recognition system that can accurately recognize gestures outside of the predefined gesture set with zero extra training effort. The key insight of XGest is to build a knowledge transfer framework between different gesture datasets. Specifically, we design a novel deep neural network to embed gestures into a high-dimensional Euclidean space. Several techniques are designed to tackle the spatial resolution limits imposed by RF hardware and the specular reflection effect of RF signals in this model. We implement XGest on a commodity mmWave device, and extensive experiments have demonstrated the significant recognition performance.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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 · cited 1×
“CNN [13] is a widely used network component in image classification tasks, where the structure of image data are spatially correlated.”
BackgroundEmpirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
Chung, J., Gulcehre, C., Cho, K., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1412.3555 · Cited by 11,008 · cited 9×
“On each branch, a GRU [9] layer is applied to exploit the temporal feature: where Ω 2 is the GRU parameters.”
MethodInfluentialFrom RSSI to CSI: Indoor localization via channel response
Yang, Z., Zhou, Z., & Liu, Y. (2013) · ACM Computing Surveys · doi:10.1145/2543581.2543592 · Cited by 1,229
Self-citationUnderstanding and Modeling of WiFi Signal Based Human Activity Recognition
Wang, W., Liu, A. X., Shahzad, M., Ling, K., & Lu, S. (2015) · Proceedings of the 21st Annual International Conference on Mobile Computing and Networking · doi:10.1145/2789168.2790093 · Cited by 1,110 · cited 4×
“For example, early attempts have lever-aged WiFi signals (e.g., CARM [37] and E-Eyes [38]) and mmWave radios (e.g., RF-Action [14] and [17]) for activity recognition.”
BackgroundMethodInfluentialSoli: ubiquitous gesture sensing with millimeter wave radar
Lien, J., Gillian, N., Karagozler, M. E., Amihood, P., Schwesig, C., Olson, E., Raja, H., & Poupyrev, I. (2016) · ACM Transactions on Graphics · doi:10.1145/2897824.2925953 · Cited by 957 · cited 1×
“Recently, some works have exploited the feasibility of gesture recognition with mmWave signals [14, 15, 39].”
BackgroundE-eyes: device-free location-oriented activity identification using fine-grained WiFi signatures
Wang, Y., Liu, J., Chen, Y., Gruteser, M., Yang, J., & Liu, H. (2014) · Proceedings of the 20th annual international conference on Mobile computing and networking · doi:10.1145/2639108.2639143 · Cited by 817 · cited 3×
“For example, early attempts have lever-aged WiFi signals (e.g., CARM [37] and E-Eyes [38]) and mmWave radios (e.g., RF-Action [14] and [17]) for activity recognition.”
BackgroundTowards Environment Independent Device Free Human Activity Recognition
Jiang, W., Miao, C., Ma, F., Yao, S., Wang, Y., Yuan, Y., Xue, H., Song, C., Ma, X., Koutsonikolas, D., et al. (2018) · Proceedings of the 24th Annual International Conference on Mobile Computing and Networking · doi:10.1145/3241539.3241548 · Cited by 573 · cited 2×
“To validate the effectiveness of this decomposition, we compare XGest with two state-of-the-art works: EI [11] and Widar3.”
BackgroundMethodZero-Effort Cross-Domain Gesture Recognition with Wi-Fi
Zheng, Y., Zhang, Y., Qian, K., Zhang, G., Liu, Y., Wu, C., & Yang, Z. (2019) · Proceedings of the 17th Annual International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3307334.3326081 · Cited by 557 · cited 2×
“0 [48], we propose an algorithm to transform the Doppler frequency spectrum into the body-coordinate -velocity spectrum (BVP), which is theoretically irrelevant to location and orientation factors.”
BackgroundMethodSelf-citationCapturing the human figure through a wall
Adib, F., Hsu, C.-Y., Mao, H., Katabi, D., & Durand, F. (2015) · ACM Transactions on Graphics · doi:10.1145/2816795.2818072 · Cited by 443 · cited 1×
“Some existing works [6, 47] also identify this specular reflection effect and resolve it by exploiting the sufficient spatial resolution enabled by well-designed antenna arrays.”
BackgroundWidar2.0: Passive Human Tracking with a Single Wi-Fi Link
Qian, K., Wu, C., Zhang, Y., Zhang, G., Yang, Z., & Liu, Y. (2018) · Proceedings of the 16th Annual International Conference on Mobile Systems, Applications, and Services · doi:10.1145/3210240.3210314 · Cited by 438
Self-citation
Research works printed but not linked: 7
- Heba Abdelnasser , Moustafa Youssef , and Khaled A . Harras . 2015 . WiGest: A ubiquitous wifi-based gesture recognition system. In Proceedings of IEEE INFOCOM. Heba Abdelnasser, Moustafa Youssef, and Khaled A. Harras. 2015. WiGest: A ubiquitous wifi-based gesture recognition system. In Proceedings of IEEE INFOCOM.
- Nancy Berryman Reese and William D . Bandy . 2016 . Joint Range of Motion and Muscle Length Testing (3rd ed.).Elsevier Health Sciences . Nancy Berryman Reese and William D. Bandy. 2016. Joint Range of Motion and Muscle Length Testing (3rd ed.).Elsevier Health Sciences.
- Dushyant Mehta , Srinath Sridhar , Oleksandr Sotnychenko , Helge Rhodin , Mohammad Shafiei , Hans-Peter Seidel , Weipeng Xu , Dan Casas , and Christian Theobalt . 2017 . VNect: Real-time 3D human pose estimation with a single RGB camera. ACM Transactions on Graphics 36 (July 2017) , Article 44 , 1 – 14 . Dushyant Mehta, Srinath Sridhar, Oleksandr Sotnychenko, Helge Rhodin, Mohammad Shafiei, Hans-Peter Seidel, Weipeng Xu, Dan Casas, and Christian Theobalt. 2017. VNect: Real-time 3D human pose estimation with a single RGB camera. ACM Transactions on Graphics 36 (July 2017), Article 44, 1–14.
- Joel C. Perry and Jacob Rosen . 2006 . Design of a 7 degree-of-freedom upper-limb powered exoskeleton. In Proceedings of IEEE BioRob . Joel C. Perry and Jacob Rosen. 2006. Design of a 7 degree-of-freedom upper-limb powered exoskeleton. In Proceedings of IEEE BioRob.
- Ashish Vaswani , Noam Shazeer , Niki Parmar , Jakob Uszkoreit , Llion Jones , Aidan N. Gomez , Lukasz Kaiser , and Illia Polosukhin . 2017 . Attention is all you need . In Proceedings of NIPS. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Proceedings of NIPS.
- Chenshu Wu , Feng Zhang , Yusen Fan , and K. J. Ray Liu . 2019. RF-based inertial measurement . In Proceedings of ACM SIGCOMM. 117–129 . Chenshu Wu, Feng Zhang, Yusen Fan, and K. J. Ray Liu. 2019. RF-based inertial measurement. In Proceedings of ACM SIGCOMM. 117–129.
- Yi Zhang , Yue Zheng , Guidong Zhang , Kun Qian , Chen Qian , and Zheng Yang . 2020 . GaitID: Robust wi-fi based gait recognition . In Proceedings of Springer WASA. Yi Zhang, Yue Zheng, Guidong Zhang, Kun Qian, Chen Qian, and Zheng Yang. 2020. GaitID: Robust wi-fi based gait recognition. In Proceedings of Springer WASA.
Cited by
Xu, H., Zhou, P., Tan, R., Li, M., & Shen, G. (2021) · Proceedings of the 19th ACM Conference on Embedded Networked Sensor Systems · doi:10.1145/3485730.3485937 · Cited by 150
Radar-Based Air-Writing Gesture Recognition Using a Novel Multistream CNN Approach
Ahmed, S., Kim, W., Park, J., & Cho, S. H. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3189395 · Cited by 95
“For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ through radar has widely been explored [2], [13].”
BackgroundInfluentialMulti-Graph Heterogeneous Interaction Fusion for Social Recommendation
Zhang, C., Wang, Y., Zhu, L., Song, J., & Yin, H. (2021) · ACM Transactions on Information Systems · doi:10.1145/3466641 · Cited by 69
XRF55: A Radio Frequency Dataset for Human Indoor Action Analysis
Wang, F., Lv, Y., Zhu, M., Ding, H., & Han, J. (2024) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3643543 · Cited by 112
“Besides, inspired by [24, 72], we leverage a pre-trained BERT [10] to output semantic embedding over the names of action classes, which we further regard as classification anchors.”
MethodRethinking Fall Detection With Wi-Fi
Yang, Z., Zhang, Y., & Zhang, Q. (2022) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2022.3188779 · Cited by 63
“[43], [44], [45] extract Doppler shifts or higher level features from wireless signals to perform human gesture recognition.”
BackgroundSelf-citationCDFi: Cross-Domain Action Recognition Using WiFi Signals
Sheng, B., Han, R., Cai, H., Xiao, F., Gui, L., & Guo, Z. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2023.3348939 · Cited by 48
“In the future, we can extend the CDFi system into zero-shot learning framework via attribute embeddings [57], [58].”
MethodDeep Ensemble Learning for Human Activity Recognition Using Wearable Sensors via Filter Activation
Huang, W., Zhang, L., Wang, S., Wu, H., & Song, A. (2022) · ACM Transactions on Embedded Computing Systems · doi:10.1145/3551486 · Cited by 53
“During the past decade, with rapid development of sensor and Internet of Things technology, human activity recognition ( HAR ) [1–3] using wearable sensors such as accelerometer and gyroscope has become a new research trend.”
BackgroundSelf-citationMetaFormer: Domain-Adaptive WiFi Sensing with Only One Labelled Target Sample
Sheng, B., Han, R., Xiao, F., Guo, Z., & Gui, L. (2024) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3643550 · Cited by 31
“Considering the large scale of vision data, an intuitive idea is to utilize available images or videos to assist wireless sensing [39, 52].”
MethodA Reputation-Based Collaborative User Recruitment Algorithm in Edge-Aided Mobile Crowdsensing
Liu, Y., Li, Y., Cheng, W., Wang, W., & Yang, J. (2023) · Applied Sciences · doi:10.3390/app13106040 · Cited by 13
Liao, P., Wang, X., An, L., Mao, S., Zhao, T., & Yang, C. (2023) · ACM Transactions on Sensor Networks · doi:10.1145/3614096 · Cited by 20
Cross-Domain Gesture Sequence Recognition for Two-Player Exergames using COTS mmWave Radar
Akbar, A. J., Sheng, Z., Zhang, Q., & Wang, D. (2023) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3626477 · Cited by 10
“Some of them focus on gesture recognition [26, 41, 68], object tracking [30, 59], user identification [61] and health care [19].”
BackgroundMethodNiaz, F., Zhang, J., Yang, Z., Khalid, M., & Niaz, A. (2024) · ACM Transactions on Sensor Networks · doi:10.1145/3694970 · Cited by 6
“• To the best of our knowledge, this is the pioneering effort to explore the potential of utilizing mmWave RF • We design a suite of advanced algorithms to extract meaningful data patterns from the RF signal snapshots and”
Self-citationRF-Eye: Commodity RFID Can Know What You Write and Who You Are Wherever You Are
Feng, Y., Huang, J., Zhang, Y., Zhang, X., Li, M., Zhang, F., Zheng, T., Li, A., Dong, M., & Liu, Z. (2025) · ACM Transactions on Sensor Networks · doi:10.1145/3737701 · Cited by 5
“XGest [52] is a cross-label gesture recognition system that leverages a novel knowledge transfer framework for accurate gesture recognition without additional training.”
Self-citationTowards Customizable Foundation Models for Human Activity Recognition with Wearable Devices
Qiu, M., Weng, C., Fan, M., & Wu, K. (2025) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3749479 · Cited by 4
“Additionally, HAR systems incorporate vision-based [32], wireless-based [16, 21, 50, 61, 66, 67], and mobile-based [17, 24, 27, 38, 40, 41, 43, 55, 57, 59] methods.”
Hybrid Deep Learning-Based Fault Detection for IoT Sensors
Rajendran, A., Dhasinamurthi, A., & Uthayakumar, G. S. (2025) · International Conference on Advances in Modern Age Technologies for Health and Engineering Science (AMATHE) · doi:10.1109/amathe65477.2025.11080927 · Cited by 3
Zhang, X., Yue, C., Zhang, Y., Liang, K., & Li, Y. (2025) · IEEE Access · doi:10.1109/access.2025.3572406 · Cited by 3
“Zhang et al. [37] developed an RFID-based neural network model called XGest, which embeds gestures into high-dimensional Euclidean space for cross-label gesture recognition.”
Self-citationmmZeAR: Zero-Effort Cross-Category Action Recognition With mmWave Radar
Sheng, B., Li, J., Cai, H., Zuo, Y., Lü, L., & Xiao, F. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3573168 · Cited by 4
“(cid:2) Xgest [59] inputs the range-velocity spectrogram series into the CNN-LSTM structure to learn spatial-temporal feature representations for gestures.”
InfluentialMultiGes: Real-Time Multi-Target Gesture Recognition for ISAC-Driven Human–Computer Interaction
Guo, Z., Wang, D., Cui, K., Gui, L., Ye, N., & Xiao, F. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3609762 · Cited by 3
“XGest [17] combines deep neural networks with mmWave technology, enabling cross-domain gesture recognition.”
Sun, Z., Liao, G., Lv, Z., Zhao, G., & Li, C. (2021) · Security and Communication Networks · doi:10.1155/2021/1094044 · Cited by 2
RetractedHicheri, R., Avazov, N., Muaaz, M., & Pätzold, M. (2025) · IEEE Access · doi:10.1109/access.2025.3646360 · Cited by 3
“Moreover, an RF-based gesture recognition approach that facilitates knowledge transfer between different datasets without requiring additional training efforts was presented in [29].”
iMe: Incremental, Memory-Efficient Hand Gesture Recognition System With Wi-Fi
Chang, L., Xie, G., Zhang, Y., Yang, X., & Wang, J. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3439376 · Cited by 1
“RF-based gesture recognition [29], [30], [31], on the other hand, does not require wearing specialized devices and is not limited by lighting or similar conditions, making it widely studied and of great interest.”
Self-citationMultiscale Feature Fusion for Gesture Recognition Using Commodity Millimeter-Wave Radar
Li, L., Bai, W., & Han, C. (2024) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2024.056073 · Cited by 1
Investigation of Input Modalities Based on a Spatial Region Array for Hand-Gesture Interfaces
Wu, H., Han, Y., Zhou, Y., Zhang, X., Yin, J., & Wang, S. (2021) · Electronics · doi:10.3390/electronics10243078 · Cited by 2
“A novel deep neural network is designed in [7], which embeds gestures in the high-dimensional Euclidean space.”
BackgroundBuilding generalizable deep learning solutions for mobile sensing
Xu, H. (2024) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/174809 · Cited by 1
I Sense You Fast: Simultaneous Action and Identity Inference by Slimming Multi-Branch RadarNet
Sheng, B., Bao, Y., Cai, H., Gui, L., & Xiao, F. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3570757
He, Y., Fang, Y., Zhang, Y., & Zhou, D. (2025) · Journal of Advanced Computational Intelligence and Intelligent Informatics · doi:10.20965/jaciii.2025.p1172
Self-citationRTS: Towards Underground Root Tuber Sensing via RF Sensor Networks
Wang, T., Zhao, Y., Huang, Z., Liu, J., & Zhong, Y. (2026) · ACM Transactions on Sensor Networks · doi:10.1145/3796515
“Despite the absence of existing work using RF tomography for sensing underground biomass, some previous RF-based works leveraging radio signals have been proposed for human sensing [39, 81], object localizing [52, 60], fruit ripeness detecting [2], and underground root sensing [1, 42, 43].”
Pendyala, V. S., & Patil, M. (2026) · International Conference on Computational Intelligence in Data Science (ICCIDS) · doi:10.1109/iccids69108.2026.11407639
Few-Shot Unseen Gestures Recognition Via Enhancing Multimodal Protonet with Dive Before Fly Fusion
Yan, Y., Lv, N., Du, W., Chen, H., Zhang, Y., & Yang, H. (2026) · Proceedings - International Conference on Image Processing · doi:10.1109/icip61757.2026.11630183
Self-citation
Metadata
- Authors
- Zhang, Y., Yang, Z., Zhang, G., Wu, C., & Zhang, L.
- Affiliations
- Tsinghua University (CN)
- University of Maryland, College Park (US)
- Hefei University of Technology (CN)
- Venue
- ACM Transactions on Sensor Networks
- Published
- 2021-09-03
- Volume
- 17
- Issue
- 4
- Pages
- 1-23
- ISSN
- 1550-4859 (ISSN-L)
- 1550-4867
- OpenAlex
- W3198846710
- Cited by
- 29
- References
- 38
- Citations by year
- 2026: 3 · 2025: 9 · 2024: 7 · 2023: 3 · 2022: 3 · 2021: 4
- FWCI
- 2.61
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- ACM Transactions on Sensor Networks (publishedVersion)
- Funders
- National Key R&D Program of China (2018AAA0901200)
- NSFC (61832010, 61632008, 61872081, 61632013, and 61972131)
- Topics
- Physical Sciences › Computer Science › Human-Computer Interaction › Hand Gesture Recognition Systems
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Indoor and Outdoor Localization Technologies
- Keywords
- gesture recognition, radio frequency signals, specular reflection, spatial resolution limits, zero-shot learning
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.