# SHREC 2022 Track on Online Detection of Heterogeneous Gestures

**Authors:** Ariel Caputo, Marco Emporio, Andrea Giachetti, Marco Cristani, Guido Borghi, Andrea D'Eusanio, Minh-Quan Le, Hai-Dang Nguyen, Minh-Triet Tran, F. Ambellan, M. Hanik, E. Nava-Yazdani, C. von Tycowicz  
**arXiv:** [2207.06706](https://arxiv.org/abs/2207.06706)  
**Date:** 2022-07-14  
**Summary page:** https://www.summarizepaper.com/en/paper/2207.06706v1/

## Key points

- Contest aimed to assess methods for recognizing diverse gestures online using sequences of 3D hand poses
- Dataset consisted of continuous hand tracking data collected in a realistic mixed reality setting using Hololens 2 finger tracking system
- Evaluation included detection performance, latency, and false positives to assess practicality of proposed algorithms
- Need for further research to minimize recognition errors while affirming low computational cost
- Authors involved in the research include Ariel Caputo, Marco Emporio, Andrea Giachetti, and others
- Study emphasizes advancing gesture recognition technology for practical applications in mixed reality environments

## Summary

, , , , 

The paper "SHREC 2022 Track on Online Detection of Heterogeneous Gestures" presents the results of a contest that aimed to assess methods for recognizing diverse gestures online using sequences of 3D hand poses. The objective was to detect gestures from a set of 16 classes with varying pose and motion characteristics. The dataset consisted of continuous hand tracking data collected in a realistic mixed reality setting using the Hololens 2 finger tracking system, where the target gestures were interspersed with non-significant movements. Along with evaluating detection performance, the study also considered factors such as latency and false positives to gauge the practicality of implementing proposed algorithms in interactive tools. The outcomes highlighted the need for further research to minimize recognition errors while affirming that the computational cost associated with the algorithms was sufficiently low. Authors involved in this research include Ariel Caputo, Marco Emporio, Andrea Giachetti, Marco Cristani, Guido Borghi, Andrea D'Eusanio, Minh-Quan Le, Hai-Dang Nguyen, Minh-Triet Tran, F. Ambellan, M. Hanik, E. Nava-Yazdani and C. von Tycowicz. The paper has been accepted for publication in the Computer & Graphics journal. Overall, this study underscores the importance of advancing gesture recognition technology for practical applications in mixed reality environments and highlights avenues for future research to enhance accuracy while maintaining computational efficiency.