Depth Super-Resolution from Explicit and Implicit High-Frequency Features

Xin QiaoChenyang GeYoumin ZhangYanhui ZhouFabio TosiMatteo PoggiStefano Mattoccia

Xin QiaoChenyang GeYoumin Zhang ...+3 Stefano Mattoccia

Mar 2023

摘要

We propose a novel multi-stage depth super-resolution network, whichprogressively reconstructs high-resolution depth maps from explicit andimplicit high-frequency features. The former are extracted by an efficienttransformer processing both local and global contexts, while the latter areobtained by projecting color images into the frequency domain. Both arecombined together with depth features by means of a fusion strategy within amulti-stage and multi-scale framework. Experiments on the main benchmarks, suchas NYUv2, Middlebury, DIML and RGBDD, show that our approach outperformsexisting methods by a large margin (~20% on NYUv2 and DIML against thecontemporary work DADA, with 16x upsampling), establishing a newstate-of-the-art in the guided depth super-resolution task.

展开全部

图表提取

论文十问

参考文献

被引用

提问与回答

暂无人提供速读十问回答

被引用

笔记

问答

领域