English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

We present an unsupervised simultaneous learning framework for the task ofmonocular camera re-localization and depth estimation from unlabeled videosequences. Monocular camera re-localization refers to the task of estimatingthe absolute camera pose from an instance image in a known environment, whichhas been intensively studied for alternative localization in GPS-deniedenvironments. In recent works, camera re-localization methods are trained viasupervised learning from pairs of camera images and camera poses. In contrastto previous works, we propose a completely unsupervised learning framework forcamera re-localization and depth estimation, requiring only monocular videosequences for training. In our framework, we train two networks that estimatethe scene coordinates using directions and the depth map from each image whichare then combined to estimate the camera pose. The networks can be trainedthrough the minimization of loss functions based on our loop closed viewsynthesis. In experiments with the 7-scenes dataset, the proposed methodoutperformed the re-localization of the state-of-the-art visual SLAM,ORB-SLAM3. Our method also outperforms state-of-the-art monocular depthestimation in a trained environment.

Unsupervised Simultaneous Learning for Camera Re-Localization and Depth Estimation from Video