A stereo matching algorithm based on the improved PSMNet.

Deep learning based on a convolutional neural network (CNN) has been successfully applied to stereo matching. Compared with the traditional method, the speed and accuracy of this method have been greatly improved. However, the existing stereo matching framework based on a CNN often encounters two pr...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Zedong Huang, Jinan Gu, Jing Li, Xuefei Yu
Formato: article
Lenguaje:EN
Publicado: Public Library of Science (PLoS) 2021
Materias:
R
Q
Acceso en línea:https://doaj.org/article/318bf79f1e0f4d60a0d3ba5b0d223b02
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:318bf79f1e0f4d60a0d3ba5b0d223b02
record_format dspace
spelling oai:doaj.org-article:318bf79f1e0f4d60a0d3ba5b0d223b022021-12-02T20:17:48ZA stereo matching algorithm based on the improved PSMNet.1932-620310.1371/journal.pone.0251657https://doaj.org/article/318bf79f1e0f4d60a0d3ba5b0d223b022021-01-01T00:00:00Zhttps://doi.org/10.1371/journal.pone.0251657https://doaj.org/toc/1932-6203Deep learning based on a convolutional neural network (CNN) has been successfully applied to stereo matching. Compared with the traditional method, the speed and accuracy of this method have been greatly improved. However, the existing stereo matching framework based on a CNN often encounters two problems. First, the existing stereo matching network has many parameters, which leads to the matching running time being too long. Second, the disparity estimation is inadequate in some regions where reflections, repeated textures, and fine structures may lead to ill-posed problems. Through the lightweight improvement of the PSMNet (Pyramid Stereo Matching Network) model, the common matching effect of ill-conditioned areas such as repeated texture areas and weak texture areas is solved. In the feature extraction part, ResNeXt is introduced to learn unitary feature extraction, and the ASPP (Atrous Spatial Pyramid Pooling) module is trained to extract multiscale spatial feature information. The feature fusion module is designed to effectively fuse the feature information of different scales to construct the matching cost volume. The improved 3D CNN uses the stacked encoding and decoding structure to further regularize the matching cost volume and obtain the corresponding relationship between feature points under different parallax conditions. Finally, the disparity map is obtained by a regression. We evaluate our method on the Scene Flow, KITTI 2012, and KITTI 2015 stereo datasets. The experiments show that the proposed stereo matching network achieves a comparable prediction accuracy and much faster running speed compared with PSMNet.Zedong HuangJinan GuJing LiXuefei YuPublic Library of Science (PLoS)articleMedicineRScienceQENPLoS ONE, Vol 16, Iss 8, p e0251657 (2021)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Zedong Huang
Jinan Gu
Jing Li
Xuefei Yu
A stereo matching algorithm based on the improved PSMNet.
description Deep learning based on a convolutional neural network (CNN) has been successfully applied to stereo matching. Compared with the traditional method, the speed and accuracy of this method have been greatly improved. However, the existing stereo matching framework based on a CNN often encounters two problems. First, the existing stereo matching network has many parameters, which leads to the matching running time being too long. Second, the disparity estimation is inadequate in some regions where reflections, repeated textures, and fine structures may lead to ill-posed problems. Through the lightweight improvement of the PSMNet (Pyramid Stereo Matching Network) model, the common matching effect of ill-conditioned areas such as repeated texture areas and weak texture areas is solved. In the feature extraction part, ResNeXt is introduced to learn unitary feature extraction, and the ASPP (Atrous Spatial Pyramid Pooling) module is trained to extract multiscale spatial feature information. The feature fusion module is designed to effectively fuse the feature information of different scales to construct the matching cost volume. The improved 3D CNN uses the stacked encoding and decoding structure to further regularize the matching cost volume and obtain the corresponding relationship between feature points under different parallax conditions. Finally, the disparity map is obtained by a regression. We evaluate our method on the Scene Flow, KITTI 2012, and KITTI 2015 stereo datasets. The experiments show that the proposed stereo matching network achieves a comparable prediction accuracy and much faster running speed compared with PSMNet.
format article
author Zedong Huang
Jinan Gu
Jing Li
Xuefei Yu
author_facet Zedong Huang
Jinan Gu
Jing Li
Xuefei Yu
author_sort Zedong Huang
title A stereo matching algorithm based on the improved PSMNet.
title_short A stereo matching algorithm based on the improved PSMNet.
title_full A stereo matching algorithm based on the improved PSMNet.
title_fullStr A stereo matching algorithm based on the improved PSMNet.
title_full_unstemmed A stereo matching algorithm based on the improved PSMNet.
title_sort stereo matching algorithm based on the improved psmnet.
publisher Public Library of Science (PLoS)
publishDate 2021
url https://doaj.org/article/318bf79f1e0f4d60a0d3ba5b0d223b02
work_keys_str_mv AT zedonghuang astereomatchingalgorithmbasedontheimprovedpsmnet
AT jinangu astereomatchingalgorithmbasedontheimprovedpsmnet
AT jingli astereomatchingalgorithmbasedontheimprovedpsmnet
AT xuefeiyu astereomatchingalgorithmbasedontheimprovedpsmnet
AT zedonghuang stereomatchingalgorithmbasedontheimprovedpsmnet
AT jinangu stereomatchingalgorithmbasedontheimprovedpsmnet
AT jingli stereomatchingalgorithmbasedontheimprovedpsmnet
AT xuefeiyu stereomatchingalgorithmbasedontheimprovedpsmnet
_version_ 1718374337390575616