🎶 This project and paper has been presented at CVPR 2022 WiCV Workshop and will be published by AAAI 2023 Student Abstract and Poster Program.
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 sports referee for the coolest skateboarding competitions.
🔯 Code Link: https://pan.baidu.com/s/1x0rEGzpIb9HJj7zN5OBJ7w (key: cscy)
