Deep Convolutional Neural Network for 6-DOF Image Localization

TL;DR

This paper introduces a deep learning approach for precise 6-DOF image localization using CNN regression, leveraging synthesized data for improved accuracy in outdoor environments.

Contribution

It presents a novel method combining automatic photo synthesis from point clouds with CNN-based pose regression for 6-DOF localization.

Findings

Achieved within 1 meter and 1 degree accuracy on outdoor campus dataset.

Demonstrated robustness and accuracy of the CNN-based localization approach.

Utilized synthesized images to enhance training data quality.

Abstract

We present an accurate and robust method for six degree of freedom image localization. There are two key-points of our method, 1. automatic immense photo synthesis and labeling from point cloud model and, 2. pose estimation with deep convolutional neural networks regression. Our model can directly regresses 6-DOF camera poses from images, accurately describing where and how it was captured. We achieved an accuracy within 1 meters and 1 degree on our out-door dataset, which covers about 2 acres on our school campus.