Towards CNN map representation and compression for camera relocalisation

2017/09/15 by Luis Contreras, Contreras, Luis, Walterio Mayol-Cuevas +1
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Robotics (cs.RO) #cs.CV #cs.RO

paper · pdf · doi:10.48550/arxiv.1709.05972

Submitted to the 1st International Workshop on Deep Learning for Visual SLAM, at the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

arxiv created 2018/05/16 · arxiv updated 2018/05/17

Abstract

This paper presents a study on the use of Convolutional Neural Networks for camera relocalisation and its application to map compression. We follow state of the art visual relocalisation results and evaluate the response to different data inputs. We use a CNN map representation and introduce the notion of map compression under this paradigm by using smaller CNN architectures without sacrificing relocalisation performance. We evaluate this approach in a series of publicly available datasets over a number of CNN architectures with different sizes, both in complexity and number of layers. This formulation allows us to improve relocalisation accuracy by increasing the number of training trajectories while maintaining a constant-size CNN.

Related

• Towards CNN Map Compression for camera relocalisation
• Camera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Network
• S3E-GNN: Sparse Spatial Scene Embedding with Graph Neural Networks for Camera Relocalization
• Modelling Uncertainty in Deep Learning for Camera Relocalization
• Adapting Convolutional Neural Networks for Geographical Domain Shift