UGR Dynamic Gesture Data-Glove Dataset: An Anonymized Flexion-IMU Subset for Cross-User Gesture Recognition

This dataset provides an anonymized dynamic-gesture subset of the UGR data-glove dataset. It contains synchronized flexion and inertial measurement unit (IMU) streams collected from a multimodal data glove for dynamic hand gesture recognition. Each gesture sample is represented as a 50-frame window with 11 sensor channels, including five flexion channels, three-axis acceleration, and three-axis angular velocity. The released file contains 7,815 dynamic gesture windows, 15 anonymized gesture classes, and 18 anonymized subject identifiers. The data are intended to support research on data-glove-based dynamic gesture recognition, cross-user generalization, and lightweight personalization. The release includes the main CSV data file, sensor schema, anonymized gesture label map, subject-level summary, leave-one-subject-out evaluation protocol, checksum file, and a minimal Python loading example. Static-gesture recordings, raw videos, firmware, hardware design files, and full source code are not included due to project and privacy constraints. The dataset is released under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Users should cite the Zenodo DOI and the associated UGR manuscript when using this dataset.