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

Existing color-guided depth super-resolution (DSR) approaches require pairedRGB-D data as training samples where the RGB image is used as structuralguidance to recover the degraded depth map due to their geometrical similarity.However, the paired data may be limited or expensive to be collected in actualtesting environment. Therefore, we explore for the first time to learn thecross-modality knowledge at training stage, where both RGB and depth modalitiesare available, but test on the target dataset, where only single depth modalityexists. Our key idea is to distill the knowledge of scene structural guidancefrom RGB modality to the single DSR task without changing its networkarchitecture. Specifically, we construct an auxiliary depth estimation (DE)task that takes an RGB image as input to estimate a depth map, and train bothDSR task and DE task collaboratively to boost the performance of DSR. Uponthis, a cross-task interaction module is proposed to realize bilateral crosstask knowledge transfer. First, we design a cross-task distillation scheme thatencourages DSR and DE networks to learn from each other in a teacher-studentrole-exchanging fashion. Then, we advance a structure prediction (SP) task thatprovides extra structure regularization to help both DSR and DE networks learnmore informative structure representations for depth recovery. Extensiveexperiments demonstrate that our scheme achieves superior performance incomparison with other DSR methods.

Learning Scene Structure Guidance via Cross-Task Knowledge Transfer for Single Depth Super-Resolution