Leveraging Neural Radiance Field in Descriptor Synthesis for Keypoints Scene Coordinate Regression

arXiv:2403.10297v1 Announce Type: new
Abstract: Classical structural-based visual localization methods offer high accuracy but face trade-offs in terms of storage, speed, and privacy. A recent innovation, keypoint scene coordinate regression (KSCR) named D2S addresses these issues by leveraging graph attention networks to enhance keypoint relationships and predict their 3D coordinates using a simple multilayer perceptron (MLP). Camera pose is then determined via PnP+RANSAC, using established 2D-3D correspondences. While KSCR achieves competitive results, rivaling state-of-the-art image-retrieval methods like HLoc across multiple …

arxiv cs.cv neural radiance field regression synthesis type