[論文レビュー] Gesture Recognition from body-Worn RFID under Missing Data

Sahar Golipoor, Richard T. Brophy+3 more|arXiv (Cornell University)|Jan 22, 2026

Hand Gesture Recognition Systems被引用数 0

ひとこと要約

The paper develops a gesture recognition system using body-worn passive RFID tags, introducing missing-data handling via interpolation, imputation, and a graph-attention CNN, achieving 98.13% accuracy on 21 gestures and 89.28% with leave-one-person-out.

ABSTRACT

We explore hand-gesture recognition through the use of passive body-worn reflective tags. A data processing pipeline is proposed to address the issue of missing data. Specifically, missing information is recovered through linear and exponential interpolation and extrapolation. Furthermore, imputation and proximity-based inference are employed. We represent tags as nodes in a temporal graph, with edges formed based on correlations between received signal strength (RSS) and phase values across successive timestamps, and we train a graph-based convolutional neural network that exploits graph-based self-attention. The system outperforms state-of-the-art methods with an accuracy of 98.13% for the recognition of 21 gestures. We achieve 89.28% accuracy under leave-one-person-out cross-validation. We further investigate the contribution of various body locations on the recognition accuracy. Removing tags from the arms reduces accuracy by more than 10%, while removing the wrist tag only reduces accuracy by around 2%. Therefore, tag placements on the arms are more expressive for gesture recognition than on the wrist.

研究の動機と目的

• Motivate robust gesture recognition from body-worn RFID tags despite tag miss-detections and data loss.
• Propose a data processing pipeline including interpolation, imputation, and normalization to recover missing information.
• Introduce a graph-based neural network that uses RSS and phase correlations across tags and time for classification.
• Evaluate the system across multiple environments, distances, and subjects to analyze placement and robustness.

提案手法

• Represent eight body-worn RFID tags as nodes in a temporal graph with edges based on RSS/phase correlations across timestamps.
• Apply phase unwrapping, normalization, and Savitzky–Golay and Gaussian smoothing to denoise signals.
• Use linear and exponential interpolation to fill sparse leading/trailing zero values and zero-padding for missing samples.
• Perform within-class proximity-based imputation using Mean Euclidean Distance and spatial proximity (tag placement) to fill null dataframes.
• Construct a graph neural network with temporal-KNN graph construction and self-attention for message passing and aggregation.
• Structure input as a 4D tensor [B, T, N, D] with T=30, N=8, D=2 (RSS and phase) for graph learning.

実験結果

リサーチクエスチョン

• RQ1 Can body-worn RFID backscatter signals from multiple tags reliably distinguish 21 hand gestures under missing data conditions?
• RQ2 How do data cleaning strategies (interpolation, imputation) influence recognition performance and robustness to tag loss?
• RQ3 What is the impact of tag placement on recognition accuracy and how can a graph-based model leverage correlations across tags and time?
• RQ4 Does a graph-based self-attention CNN outperform traditional RF/RSS-based classifiers under leave-one-person-out scenarios?

主な発見

Method	Acc.	Pre.	Rec.	F1
RFC with SP	83.56	83.72	83.56	83.42
RFC with SWP	86.26	86.48	86.20	86.07
RFC with SPR	95.25	95.35	95.25	95.23
Early Fusion	83.62	84.25	83.57	83.34
Late Fusion	87.13	88.32	87.08	86.91
EUIGR	80.40	78.94	80.07	78.73
GRfid	30.34	31.40	30.16	30.22
Our model	98.13	98.19	98.13	98.13
RFC with SP	85.07	86.29	85.02	85.13
RFC with SWP	85.71	86.72	85.71	85.72
RFC with SPR	93.52	94.35	93.52	93.59
Early Fusion	81.01	86.08	81.00	81.18
Late Fusion	89.41	90.18	89.39	89.39
EUIGR	80.37	74.75	80.33	75.97
GRfid	29.39	29.14	29.34	28.89
Our model	96.82	97.88	96.80	97.02
RFC with SP	84.12	84.67	84.07	83.96
RFC with SWP	81.26	82.39	81.26	80.93
RFC with SPR	93.80	94.08	93.78	93.71
Early Fusion	91.13	94.16	91.13	91.56
Late Fusion	90.15	91.13	90.15	90.15
EUIGR	87.30	83.75	87.27	84.55
GRfid	34.43	36.30	33.87	34.33
Our model	98.41	-	-	-

• Achieves 98.13% accuracy on 21 gestures in within-user testing and 89.28% with leave-one-person-out cross-validation.
• Single-hand gesture accuracy reaches 98.27% for some gestures, with 16 of 21 gestures at 100% in within-user tests.
• Removing arm-mounted tags significantly reduces accuracy ( >10% drop when arms’ tags are removed; wrist-tag removal only ~2% drop).
• All eight tags contribute to high performance; certain tags (T4 and T8) have the largest impact when omitted.
• The proposed graph-attention framework outperforms RF-based baselines (e.g., RFC with SPR, Early/Late Fusion, EUIGR, GRfid) across three datasets at distances 3m and 1.5m.