SAIC_Cambridge-HuPBA-FBK Submission to the EPIC-Kitchens-100 Action Recognition Challenge 2021
This work addresses action recognition for video analysis, but it is incremental as it applies existing methods to a new challenge.
The paper tackled action recognition in videos by deploying an ensemble of GSF and XViT models, achieving a top-1 accuracy of 44.82% on the EPIC-Kitchens-100 dataset using only RGB data.
This report presents the technical details of our submission to the EPIC-Kitchens-100 Action Recognition Challenge 2021. To participate in the challenge we deployed spatio-temporal feature extraction and aggregation models we have developed recently: GSF and XViT. GSF is an efficient spatio-temporal feature extracting module that can be plugged into 2D CNNs for video action recognition. XViT is a convolution free video feature extractor based on transformer architecture. We design an ensemble of GSF and XViT model families with different backbones and pretraining to generate the prediction scores. Our submission, visible on the public leaderboard, achieved a top-1 action recognition accuracy of 44.82%, using only RGB.