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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Camera pose regression methods apply a single forward pass to the query imageto estimate the camera pose. As such, they offer a fast and light-weightalternative to traditional localization schemes based on image retrieval. Poseregression approaches simultaneously learn two regression tasks, aiming tojointly estimate the camera position and orientation using a single embeddingvector computed by a convolutional backbone. We propose an attention-basedapproach for pose regression, where the convolutional activation maps are usedas sequential inputs. Transformers are applied to encode the sequentialactivation maps as latent vectors, used for camera pose regression. This allowsus to pay attention to spatially-varying deep features. Using two Transformerheads, we separately focus on the features for camera position and orientation,based on how informative they are per task. Our proposed approach is shown tocompare favorably to contemporary pose regressors schemes and achievesstate-of-the-art accuracy across multiple outdoor and indoor benchmarks. Inparticular, to the best of our knowledge, our approach is the only method toattain sub-meter average accuracy across outdoor scenes. We make our codepublicly available from here.

Paying Attention to Activation Maps in Camera Pose Regression