ace and ace-g
ACE-G is a generalized successor to ACE that improves upon the original scene coordinate regression approach for visual relocalization through architectural enhancements and pre-training schemes.
About ace
nianticlabs/ace
[CVPR 2023 - Highlight] Accelerated Coordinate Encoding (ACE): Learning to Relocalize in Minutes using RGB and Poses
This project helps roboticists and augmented reality (AR) developers quickly teach a computer vision system to understand its precise location within a new environment. By inputting RGB images and their corresponding camera poses (locations and orientations), it generates a compact 'scene map' that enables fast and accurate re-localization. This tool is for professionals building navigation systems for robots or immersive AR experiences.
About ace-g
nianticspatial/ace-g
[ICCV 2025] ACE-G is an architecture and pre-training scheme to improve generalization for scene coordinate regression-based visual relocalization.
This project helps computer vision practitioners accurately determine the precise location and orientation (pose) of a camera within a known environment by identifying specific points in images and matching them to 3D coordinates in a digital map. You provide a set of images of a space, and the system creates a map that can then be used to find the camera pose of new images within that space. This is ideal for robotics engineers, augmented reality developers, or anyone needing highly precise camera tracking.
Scores updated daily from GitHub, PyPI, and npm data. How scores work