Monocular Depth Estimation Primed by Salient Point Detection and Normalized Hessian Loss

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F21%3A00364661" target="_blank" >RIV/68407700:21230/21:00364661 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1109/3DV53792.2021.00033" target="_blank" >https://doi.org/10.1109/3DV53792.2021.00033</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/3DV53792.2021.00033" target="_blank" >10.1109/3DV53792.2021.00033</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Monocular Depth Estimation Primed by Salient Point Detection and Normalized Hessian Loss

Popis výsledku v původním jazyce

Deep neural networks have recently thrived on single image depth estimation. That being said, current developments on this topic highlight an apparent compromise between accuracy and network size. This work proposes an accurate and lightweight framework for monocular depth estimation based on a self-attention mechanism stemming from salient point detection. Specifically, we utilize a sparse set of keypoints to train a FuSaNet model that consists of two major components: Fusion-Net and Saliency-Net. In addition, we introduce a normalized Hessian loss term invariant to scaling and shear along the depth direction, which is shown to substantially improve the accuracy. The proposed method achieves state-of-the-art results on NYU-Depth-v2 and KITTI while using 3.1-38.4 times smaller model in terms of the number of parameters than baseline approaches. Experiments on the SUN-RGBD further demonstrate the generalizability of the proposed method.

Název v anglickém jazyce

Monocular Depth Estimation Primed by Salient Point Detection and Normalized Hessian Loss

Popis výsledku anglicky

Deep neural networks have recently thrived on single image depth estimation. That being said, current developments on this topic highlight an apparent compromise between accuracy and network size. This work proposes an accurate and lightweight framework for monocular depth estimation based on a self-attention mechanism stemming from salient point detection. Specifically, we utilize a sparse set of keypoints to train a FuSaNet model that consists of two major components: Fusion-Net and Saliency-Net. In addition, we introduce a normalized Hessian loss term invariant to scaling and shear along the depth direction, which is shown to substantially improve the accuracy. The proposed method achieves state-of-the-art results on NYU-Depth-v2 and KITTI while using 3.1-38.4 times smaller model in terms of the number of parameters than baseline approaches. Experiments on the SUN-RGBD further demonstrate the generalizability of the proposed method.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/EF16_019%2F0000765" target="_blank" >EF16_019/0000765: Výzkumné centrum informatiky</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

3DV 2021: Proceedings of the International Conference on 3D Vision

ISBN

978-1-6654-2688-6

ISSN

2378-3826

e-ISSN

2475-7888

Počet stran výsledku

11

Strana od-do

228-238

Název nakladatele

IEEE Computer Soc.

Místo vydání

Los Alamitos, CA

Místo konání akce

Virtual

Datum konání akce

1. 12. 2021

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

000786496000023