Salvato in:
Dettagli Bibliografici
Autori principali: Nuzhdin, Anton, Nagaev, Alexander, Sautin, Alexander, Kapitanov, Alexander, Kvanchiani, Karina
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2412.01508
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913594364395520
author Nuzhdin, Anton
Nagaev, Alexander
Sautin, Alexander
Kapitanov, Alexander
Kvanchiani, Karina
author_facet Nuzhdin, Anton
Nagaev, Alexander
Sautin, Alexander
Kapitanov, Alexander
Kvanchiani, Karina
contents This paper proposes the second version of the widespread Hand Gesture Recognition dataset HaGRID -- HaGRIDv2. We cover 15 new gestures with conversation and control functions, including two-handed ones. Building on the foundational concepts proposed by HaGRID's authors, we implemented the dynamic gesture recognition algorithm and further enhanced it by adding three new groups of manipulation gestures. The ``no gesture" class was diversified by adding samples of natural hand movements, which allowed us to minimize false positives by 6 times. Combining extra samples with HaGRID, the received version outperforms the original in pre-training models for gesture-related tasks. Besides, we achieved the best generalization ability among gesture and hand detection datasets. In addition, the second version enhances the quality of the gestures generated by the diffusion model. HaGRIDv2, pre-trained models, and a dynamic gesture recognition algorithm are publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2412_01508
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle HaGRIDv2: 1M Images for Static and Dynamic Hand Gesture Recognition
Nuzhdin, Anton
Nagaev, Alexander
Sautin, Alexander
Kapitanov, Alexander
Kvanchiani, Karina
Computer Vision and Pattern Recognition
This paper proposes the second version of the widespread Hand Gesture Recognition dataset HaGRID -- HaGRIDv2. We cover 15 new gestures with conversation and control functions, including two-handed ones. Building on the foundational concepts proposed by HaGRID's authors, we implemented the dynamic gesture recognition algorithm and further enhanced it by adding three new groups of manipulation gestures. The ``no gesture" class was diversified by adding samples of natural hand movements, which allowed us to minimize false positives by 6 times. Combining extra samples with HaGRID, the received version outperforms the original in pre-training models for gesture-related tasks. Besides, we achieved the best generalization ability among gesture and hand detection datasets. In addition, the second version enhances the quality of the gestures generated by the diffusion model. HaGRIDv2, pre-trained models, and a dynamic gesture recognition algorithm are publicly available.
title HaGRIDv2: 1M Images for Static and Dynamic Hand Gesture Recognition
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.01508