Skip to Main content Skip to Navigation
Journal articles

Photometric Depth Super-Resolution

Abstract : This study explores the use of photometric techniques (shape-from-shading and uncalibrated photometric stereo) for upsampling the low-resolution depth map from an RGB-D sensor to the higher resolution of the companion RGB image. A single-shot variational approach is first put forward, which is effective as long as the target's reflectance is piecewise-constant. It is then shown that this dependency upon a specific reflectance model can be relaxed by focusing on a specific class of objects (e.g., faces), and delegate reflectance estimation to a deep neural network. A multi-shot strategy based on randomly varying lighting conditions is eventually discussed. It requires no training or prior on the reflectance, yet this comes at the price of a dedicated acquisition setup. Both quantitative and qualitative evaluations illustrate the effectiveness of the proposed methods on synthetic and real-world scenarios.
Document type :
Journal articles
Complete list of metadata

Cited literature [104 references]  Display  Hide  Download
Contributor : Yvain Queau Connect in order to contact the contributor
Submitted on : Monday, June 3, 2019 - 11:39:50 AM
Last modification on : Saturday, June 25, 2022 - 9:53:29 AM


Files produced by the author(s)



Bjoern Haefner, Songyou Peng, Alok Verma, Yvain Quéau, Daniel Cremers. Photometric Depth Super-Resolution. IEEE Transactions on Pattern Analysis and Machine Intelligence, Institute of Electrical and Electronics Engineers, 2020, 42 (10), pp.2453--2464. ⟨10.1109/TPAMI.2019.2923621⟩. ⟨hal-02145726⟩



Record views


Files downloads