GCNDepth: Self-supervised monocular depth estimation based on graph convolutional network

Autor segons l'article: Masoumian, Armin; Rashwan, Hatem A; Abdulwahab, Saddam; Cristiano, Julian; Asif, M Salman; Puig, Domenec

Departament: Enginyeria Informàtica i Matemàtiques

Autor/s de la URV: Abdellatif Fatahallah Ibrahim Mahmoud, Hatem / Abdulwahab, Saddam Abdulrhman Hamed / CRISTIANO RODRÍGUEZ, JULIÁN EFRÉN / Masoumian, Armin / Puig Valls, Domènec Savi

Paraules clau: Deep learning; Graph convolutional network; Monocular depth estimation; Self-supervision

Resum: Depth estimation is a challenging task of 3D reconstruction to enhance the accuracy sensing of environment awareness. This work brings a new solution with improvements, which increases the quantitative and qualitative understanding of depth maps compared to existing methods. Recently, convolutional neural networks (CNN) have demonstrated their extraordinary ability to estimate depth maps from monocular videos. However, traditional CNN does not support a topological structure, and they can work only on regular image regions with determined sizes and weights. On the other hand, graph convolutional networks (GCN) can handle the convolution of non-Euclidean data, and they can be applied to irregular image regions within a topological structure. Therefore, to preserve object geometric appearances and objects locations in the scene, in this work, we aim to exploit GCN for a self-supervised monocular depth estimation model. Our model consists of two parallel auto-encoder networks: the first is an auto-encoder that will depend on ResNet-50 and extract the feature from the input image and on multi-scale GCN to estimate the depth map. In turn, the second network will be used to estimate the ego-motion vector (i.e., 3D pose) between two consecutive frames based on ResNet-18. The estimated 3D pose and depth map will be used to construct the target image. A combination of loss functions related to photometric, reprojection, and smoothness is used to cope with bad depth prediction and preserve the discontinuities of the objects. Our method and performance are improved quantitatively and qualitatively. In particular, our method provided comparable and promising results with a high prediction accuracy of 89% on the publicly available KITTI dataset. Our method also offers 40% reduction in the number of trainable parameters compared to the state of the art solutions.In addition, we tested our trained model with Make3D dataset to evaluate the trained model on a new dataset with low resolution images. The source code is publicly available at (https://github.com/ArminMasoumian/GCNDepth.git)

Àrees temàtiques: Artificial intelligence; Astronomia / física; Biotecnología; Ciência da computação; Ciências agrárias i; Ciências ambientais; Ciências biológicas i; Ciências biológicas ii; Cognitive neuroscience; Computer science applications; Computer science, artificial intelligence; Direito; Educação; Engenharias i; Engenharias ii; Engenharias iii; Engenharias iv; Geociências; Interdisciplinar; Matemática / probabilidade e estatística; Medicina i; Medicina ii; Psicología; Química

Accès a la llicència d'ús: https://creativecommons.org/licenses/by/3.0/es/

Adreça de correu electrònic de l'autor: domenec.puig@urv.cat; saddam.abdulwahab@urv.cat; armin.masoumian@estudiants.urv.cat; armin.masoumian@estudiants.urv.cat; hatem.abdellatif@urv.cat; saddam.abdulwahab@urv.cat

Data d'alta del registre: 2024-09-21

Versió de l'article dipositat: info:eu-repo/semantics/publishedVersion

Enllaç font original: https://www.sciencedirect.com/science/article/abs/pii/S0925231222013601

Referència a l'article segons font original: Neurocomputing. 517 81-92

Referència de l'ítem segons les normes APA: Masoumian, Armin; Rashwan, Hatem A; Abdulwahab, Saddam; Cristiano, Julian; Asif, M Salman; Puig, Domenec (2023). GCNDepth: Self-supervised monocular depth estimation based on graph convolutional network. Neurocomputing, 517(), 81-92. DOI: 10.1016/j.neucom.2022.10.073

URL Document de llicència: https://repositori.urv.cat/ca/proteccio-de-dades/

DOI de l'article: 10.1016/j.neucom.2022.10.073

Entitat: Universitat Rovira i Virgili

Any de publicació de la revista: 2023

Tipus de publicació: Journal Publications