Abstract
Modeling sparse and dense image matching within a unified functional model has recently attracted increasing research interest.However, existing efforts mainly focus on improving matching accuracy while ignoring its efficiency, which is crucial for real-world applications.
Pipeline of ECO-TR
In this paper, we propose an efficient structure named Efficient Correspondence Transformer ECO-TR by finding correspondences in a coarse-to-fine manner, which significantly improves the efficiency of functional model method. To achieve this, multiple transformer blocks are stage-wisely connected to gradually refine the predicted coordinates upon a shared multi-scale feature extraction network.All the correspondences are predicted within a single feed-forward pass, given a pair of images and for arbitrary query coordinates.We further propose an adaptive query-clustering strategy and an uncertainty-based outlier detection module to cooperate with the proposed framework for faster and better predictions.
Comparison of the inference time between the proposed ECO-TR and COTR
Experimental comparisons on various sparse and dense matching tasks demonstrate the superiority of our method on both efficiency and effectiveness against existing functional matching methods. are evaluated on a single Tesla V100 GPU.
