CARLOR @ Ego4D Step Grounding Challenge: Bayesian temporal-order priors for test time refinement
This addresses the challenge of step grounding in egocentric videos for applications like activity understanding, though it appears incremental as it refines an existing model with a Bayesian prior.
The paper tackles the problem of locating temporal boundaries of activities in lengthy, untrimmed egocentric videos based on natural language descriptions, achieving state-of-the-art results with a 35.18 Recall Top-1 at 0.3 IoU and 20.48 Recall Top-1 at 0.5 IoU on the Ego4D Goal-Step dataset.
The goal of the Step Grounding task is to locate temporal boundaries of activities based on natural language descriptions. This technical report introduces a Bayesian-VSLNet to address the challenge of identifying such temporal segments in lengthy, untrimmed egocentric videos. Our model significantly improves upon traditional models by incorporating a novel Bayesian temporal-order prior during inference, enhancing the accuracy of moment predictions. This prior adjusts for cyclic and repetitive actions within videos. Our evaluations demonstrate superior performance over existing methods, achieving state-of-the-art results on the Ego4D Goal-Step dataset with a 35.18 Recall Top-1 at 0.3 IoU and 20.48 Recall Top-1 at 0.5 IoU on the test set.