-
Quantifying Epistemic Uncertainty in Absolute Pose Regression
Authors:
Fereidoon Zangeneh,
Amit Dekel,
Alessandro Pieropan,
Patric Jensfelt
Abstract:
Visual relocalization is the task of estimating the camera pose given an image it views. Absolute pose regression offers a solution to this task by training a neural network, directly regressing the camera pose from image features. While an attractive solution in terms of memory and compute efficiency, absolute pose regression's predictions are inaccurate and unreliable outside the training domain…
▽ More
Visual relocalization is the task of estimating the camera pose given an image it views. Absolute pose regression offers a solution to this task by training a neural network, directly regressing the camera pose from image features. While an attractive solution in terms of memory and compute efficiency, absolute pose regression's predictions are inaccurate and unreliable outside the training domain. In this work, we propose a novel method for quantifying the epistemic uncertainty of an absolute pose regression model by estimating the likelihood of observations within a variational framework. Beyond providing a measure of confidence in predictions, our approach offers a unified model that also handles observation ambiguities, probabilistically localizing the camera in the presence of repetitive structures. Our method outperforms existing approaches in capturing the relation between uncertainty and prediction error.
△ Less
Submitted 9 April, 2025;
originally announced April 2025.
-
Conditional Variational Autoencoders for Probabilistic Pose Regression
Authors:
Fereidoon Zangeneh,
Leonard Bruns,
Amit Dekel,
Alessandro Pieropan,
Patric Jensfelt
Abstract:
Robots rely on visual relocalization to estimate their pose from camera images when they lose track. One of the challenges in visual relocalization is repetitive structures in the operation environment of the robot. This calls for probabilistic methods that support multiple hypotheses for robot's pose. We propose such a probabilistic method to predict the posterior distribution of camera poses giv…
▽ More
Robots rely on visual relocalization to estimate their pose from camera images when they lose track. One of the challenges in visual relocalization is repetitive structures in the operation environment of the robot. This calls for probabilistic methods that support multiple hypotheses for robot's pose. We propose such a probabilistic method to predict the posterior distribution of camera poses given an observed image. Our proposed training strategy results in a generative model of camera poses given an image, which can be used to draw samples from the pose posterior distribution. Our method is streamlined and well-founded in theory and outperforms existing methods on localization in presence of ambiguities.
△ Less
Submitted 7 October, 2024;
originally announced October 2024.
-
A Probabilistic Framework for Visual Localization in Ambiguous Scenes
Authors:
Fereidoon Zangeneh,
Leonard Bruns,
Amit Dekel,
Alessandro Pieropan,
Patric Jensfelt
Abstract:
Visual localization allows autonomous robots to relocalize when losing track of their pose by matching their current observation with past ones. However, ambiguous scenes pose a challenge for such systems, as repetitive structures can be viewed from many distinct, equally likely camera poses, which means it is not sufficient to produce a single best pose hypothesis. In this work, we propose a prob…
▽ More
Visual localization allows autonomous robots to relocalize when losing track of their pose by matching their current observation with past ones. However, ambiguous scenes pose a challenge for such systems, as repetitive structures can be viewed from many distinct, equally likely camera poses, which means it is not sufficient to produce a single best pose hypothesis. In this work, we propose a probabilistic framework that for a given image predicts the arbitrarily shaped posterior distribution of its camera pose. We do this via a novel formulation of camera pose regression using variational inference, which allows sampling from the predicted distribution. Our method outperforms existing methods on localization in ambiguous scenes. Code and data will be released at https://github.com/efreidun/vapor.
△ Less
Submitted 5 January, 2023;
originally announced January 2023.
-
SDF-based RGB-D Camera Tracking in Neural Scene Representations
Authors:
Leonard Bruns,
Fereidoon Zangeneh,
Patric Jensfelt
Abstract:
We consider the problem of tracking the 6D pose of a moving RGB-D camera in a neural scene representation. Different such representations have recently emerged, and we investigate the suitability of them for the task of camera tracking. In particular, we propose to track an RGB-D camera using a signed distance field-based representation and show that compared to density-based representations, trac…
▽ More
We consider the problem of tracking the 6D pose of a moving RGB-D camera in a neural scene representation. Different such representations have recently emerged, and we investigate the suitability of them for the task of camera tracking. In particular, we propose to track an RGB-D camera using a signed distance field-based representation and show that compared to density-based representations, tracking can be sped up, which enables more robust and accurate pose estimates when computation time is limited.
△ Less
Submitted 4 May, 2022;
originally announced May 2022.
