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

In this work, we tackle the essential problem of scale inconsistency forself-supervised joint depth-pose learning. Most existing methods assume that aconsistent scale of depth and pose can be learned across all input samples,which makes the learning problem harder, resulting in degraded performance andlimited generalization in indoor environments and long-sequence visual odometryapplication. To address this issue, we propose a novel system that explicitlydisentangles scale from the network estimation. Instead of relying on PoseNetarchitecture, our method recovers relative pose by directly solving fundamentalmatrix from dense optical flow correspondence and makes use of a two-viewtriangulation module to recover an up-to-scale 3D structure. Then, we align thescale of the depth prediction with the triangulated point cloud and use thetransformed depth map for depth error computation and dense reprojection check.Our whole system can be jointly trained end-to-end. Extensive experiments showthat our system not only reaches state-of-the-art performance on KITTI depthand flow estimation, but also significantly improves the generalization abilityof existing self-supervised depth-pose learning methods under a variety ofchallenging scenarios, and achieves state-of-the-art results amongself-supervised learning-based methods on KITTI Odometry and NYUv2 dataset.Furthermore, we present some interesting findings on the limitation ofPoseNet-based relative pose estimation methods in terms of generalizationability. Code is available at https://github.com/B1ueber2y/TrianFlow.

Towards Better Generalization: Joint Depth-Pose Learning without PoseNet