Impressed by the coolest skateboarding sports program from 2021 Tokyo Olympic Games, we are the first to curate the original real-world video datasets "SkateboardAI" in the wild, even self-design and implement diverse uni-modal and multi-modal video action recognition approaches to recognize different tricks accurately. For uni-modal methods, we separately apply (1) CNN and LSTM; (2) CNN and BiLSTM; (3) CNN and BiLSTM with effective attention mechanisms; (4) Transformer-based action recognition pipeline. Transferred to the multi-modal conditions, we investigated the two-stream Inflated-3D architecture on "SkateboardAI" datasets to compare its performance with uni-modal cases. In sum, our objective is developing an excellent AI sport referee for the coolest skateboarding competitions.
@article{arxiv.2311.11467,
title = {SkateboardAI: The Coolest Video Action Recognition for Skateboarding},
author = {Hanxiao Chen},
journal= {arXiv preprint arXiv:2311.11467},
year = {2024}
}
Comments
The original first-author work has been accepted and presented by CVPR 2022 WiCV Workshop (This is the long-version paper)