TY - JOUR
T1 - Single-Stage Refinement CNN for Depth Estimation in Monocular Images
AU - Rodríguez, José E.Valdez
AU - Calvo, Hiram
AU - Ón, Edgardo M.Felipe River
N1 - Publisher Copyright:
© 2020 Instituto Politecnico Nacional. All rights reserved.
PY - 2020
Y1 - 2020
N2 - Depth reconstruction from single monocular images has been a challenging task due to the complexity and the quantity of depth cues that images have. Convolutional Neural Networks (CNN) have been successfully used to reconstruct depth of general object scenes; however, proposed works use several stages of training which make this process more complex and time consuming. As we aim to build a computational efficient model, we focus on single-stage training CNN. In this paper, we propose five different models for solving this task, ranging from a simple convolutional network, to one with residual, convolutional, refinement and upsampling layers. We compare our models with the current state of the art in depth reconstruction and measure depth reconstruction error for different datasets (KITTI, NYU), obtaining improvements in both global and local error measures.
AB - Depth reconstruction from single monocular images has been a challenging task due to the complexity and the quantity of depth cues that images have. Convolutional Neural Networks (CNN) have been successfully used to reconstruct depth of general object scenes; however, proposed works use several stages of training which make this process more complex and time consuming. As we aim to build a computational efficient model, we focus on single-stage training CNN. In this paper, we propose five different models for solving this task, ranging from a simple convolutional network, to one with residual, convolutional, refinement and upsampling layers. We compare our models with the current state of the art in depth reconstruction and measure depth reconstruction error for different datasets (KITTI, NYU), obtaining improvements in both global and local error measures.
KW - Convolutional neural networks
KW - Depth reconstruction
KW - Embedded refinement layer
KW - Single stage training
KW - Stereo matching
UR - http://www.scopus.com/inward/record.url?scp=85089074705&partnerID=8YFLogxK
U2 - 10.13053/CyS-24-2-3370
DO - 10.13053/CyS-24-2-3370
M3 - Artículo
AN - SCOPUS:85089074705
SN - 1405-5546
VL - 24
SP - 439
EP - 451
JO - Computacion y Sistemas
JF - Computacion y Sistemas
IS - 2
ER -