T-Net++: Effective Permutation-Equivariance Network for Two-View Correspondence Pruning

We propose a conceptually novel, flexible, and effective framework (named T-Net++) for the task of two-view correspondence pruning. T-Net++ comprises two unique structures: the <tex-math notation="LaTeX">``-</tex-math> '' structure and the <tex-math notation="LaTeX">``|</tex-math> '' structure. The <tex-math notation="LaTeX">``-</tex-math> '' structure utilizes an iterative learning strategy to process correspondences, while the <tex-math notation="LaTeX">``|</tex-math> '' structure integrates all feature information of the <tex-math notation="LaTeX">``-</tex-math> '' structure and produces inlier weights. Moreover, within the <tex-math notation="LaTeX">``|</tex-math> '' structure, we design a new Local-Global Attention Fusion module to fully exploit valuable information obtained from concatenating features through channel-wise and spatial-wise relationships. Furthermore, we develop a Channel-Spatial Squeeze-and-Excitation module, a modified network backbone that enhances the representation ability of important channels and correspondences through the squeeze-and-excitation operation. T-Net++ not only preserves the permutation-equivariance manner for correspondence pruning, but also gathers rich contextual information, thereby enhancing the effectiveness of the network. Experimental results demonstrate that T-Net++ outperforms other state-of-the-art correspondence pruning methods on various benchmarks and excels in two extended tasks.