Scene Coordinate Regression Forests for Camera Relocalization in RGB-D Images
- 1 June 2013
- conference paper
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2930-2937
- https://doi.org/10.1109/cvpr.2013.377
Abstract
We address the problem of inferring the pose of an RGB-D camera relative to a known 3D scene, given only a single acquired image. Our approach employs a regression forest that is capable of inferring an estimate of each pixel's correspondence to 3D points in the scene's world coordinate frame. The forest uses only simple depth and RGB pixel comparison features, and does not require the computation of feature descriptors. The forest is trained to be capable of predicting correspondences at any pixel, so no interest point detectors are required. The camera pose is inferred using a robust optimization scheme. This starts with an initial set of hypothesized camera poses, constructed by applying the forest at a small fraction of image pixels. Preemptive RANSAC then iterates sampling more pixels at which to evaluate the forest, counting inliers, and refining the hypothesized poses. We evaluate on several varied scenes captured with an RGB-D camera and observe that the proposed technique achieves highly accurate relocalization and substantially out-performs two state of the art baselines.Keywords
This publication has 27 references indexed in Scilit:
- Learning to Efficiently Detect Repeatable Interest Points in Depth DataLecture Notes in Computer Science, 2012
- 6D Relocalisation for RGBD Cameras Using Synthetic View RegressionPublished by British Machine Vision Association and Society for Pattern Recognition ,2012
- KinectFusionPublished by Association for Computing Machinery (ACM) ,2011
- KinectFusion: Real-time dense surface mapping and trackingPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2011
- Automatic Relocalization and Loop Closing for Real-Time Monocular SLAMIEEE Transactions on Pattern Analysis and Machine Intelligence, 2011
- Learning Local Image DescriptorsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2007
- 3D LayoutCRF for Multi-View Object Class Recognition and SegmentationPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2007
- Keypoint recognition using randomized treesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2006
- The Layout Consistent Random Field for Recognizing and Segmenting Partially Occluded ObjectsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2006
- A solution for the best rotation to relate two sets of vectorsActa Crystallographica Section A, 1976