STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence
In order to improve the accuracy and robustness of optical flow computation under large displacements and motion occlusions, the authors present in this study a large displacement flow field estimation approach using similarity transformation‐based dense correspondence, named STDC‐Flow approach. Fir...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2020-08-01
|
| Series: | IET Computer Vision |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/iet-cvi.2019.0321 |
| _version_ | 1797684730192199680 |
|---|---|
| author | Congxuan Zhang Zhen Chen Fan Xiong Wen Liu Ming Li Liyue Ge |
| author_facet | Congxuan Zhang Zhen Chen Fan Xiong Wen Liu Ming Li Liyue Ge |
| author_sort | Congxuan Zhang |
| collection | DOAJ |
| description | In order to improve the accuracy and robustness of optical flow computation under large displacements and motion occlusions, the authors present in this study a large displacement flow field estimation approach using similarity transformation‐based dense correspondence, named STDC‐Flow approach. First, the authors compute an initial nearest‐neighbour field by using the STDC‐Flow of the consecutive two frames, and then extract the consistent regions as the robust nearest‐neighbour field and label the inconsistent regions as the occlusion areas. Second, they improve a non‐local total variation with the L1 norm optical flow model by using the occlusion information to modify the weighted median filtering optimisation. Third, they fuse the robust nearest‐neighbour field and the computed flow field of the improved variational optical flow model to construct the final flow field by using the quadratic pseudo‐boolean optimisation fusion algorithm. Finally, the authors compare the proposed STDC‐Flow method with several state‐of‐the‐art approaches including the variational and deep learning‐based optical flow models by using the MPI‐Sintel and KITTI evaluation databases. The comparison results demonstrate that the proposed STDC‐Flow method has a high accuracy for flow field computation, especially the capacity of dealing with large displacements and motion occlusions. |
| first_indexed | 2024-03-12T00:33:58Z |
| format | Article |
| id | doaj.art-41787a94b30c4954ad5ea265294715a1 |
| institution | Directory Open Access Journal |
| issn | 1751-9632 1751-9640 |
| language | English |
| last_indexed | 2024-03-12T00:33:58Z |
| publishDate | 2020-08-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Computer Vision |
| spelling | doaj.art-41787a94b30c4954ad5ea265294715a12023-09-15T10:06:16ZengWileyIET Computer Vision1751-96321751-96402020-08-0114524825810.1049/iet-cvi.2019.0321STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondenceCongxuan Zhang0Zhen Chen1Fan Xiong2Wen Liu3Ming Li4Liyue Ge5Key Laboratory of Nondestructive Testing, Ministry of Education, Nanchang Hangkong UniversityNanchangPeople's Republic of ChinaKey Laboratory of Nondestructive Testing, Ministry of Education, Nanchang Hangkong UniversityNanchangPeople's Republic of ChinaKey Laboratory of Nondestructive Testing, Ministry of Education, Nanchang Hangkong UniversityNanchangPeople's Republic of ChinaDepartment of Physical Therapy and Rehabilitation ScienceUniversity of KansasKansas CityUSAKey Laboratory of Nondestructive Testing, Ministry of Education, Nanchang Hangkong UniversityNanchangPeople's Republic of ChinaKey Laboratory of Nondestructive Testing, Ministry of Education, Nanchang Hangkong UniversityNanchangPeople's Republic of ChinaIn order to improve the accuracy and robustness of optical flow computation under large displacements and motion occlusions, the authors present in this study a large displacement flow field estimation approach using similarity transformation‐based dense correspondence, named STDC‐Flow approach. First, the authors compute an initial nearest‐neighbour field by using the STDC‐Flow of the consecutive two frames, and then extract the consistent regions as the robust nearest‐neighbour field and label the inconsistent regions as the occlusion areas. Second, they improve a non‐local total variation with the L1 norm optical flow model by using the occlusion information to modify the weighted median filtering optimisation. Third, they fuse the robust nearest‐neighbour field and the computed flow field of the improved variational optical flow model to construct the final flow field by using the quadratic pseudo‐boolean optimisation fusion algorithm. Finally, the authors compare the proposed STDC‐Flow method with several state‐of‐the‐art approaches including the variational and deep learning‐based optical flow models by using the MPI‐Sintel and KITTI evaluation databases. The comparison results demonstrate that the proposed STDC‐Flow method has a high accuracy for flow field computation, especially the capacity of dealing with large displacements and motion occlusions.https://doi.org/10.1049/iet-cvi.2019.0321motion occlusionssimilarity transformationoptical flow computationnearest-neighbour fieldL1 norm optical flow modelocclusion information |
| spellingShingle | Congxuan Zhang Zhen Chen Fan Xiong Wen Liu Ming Li Liyue Ge STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence IET Computer Vision motion occlusions similarity transformation optical flow computation nearest-neighbour field L1 norm optical flow model occlusion information |
| title | STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence |
| title_full | STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence |
| title_fullStr | STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence |
| title_full_unstemmed | STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence |
| title_short | STDC‐Flow: large displacement flow field estimation using similarity transformation‐based dense correspondence |
| title_sort | stdc flow large displacement flow field estimation using similarity transformation based dense correspondence |
| topic | motion occlusions similarity transformation optical flow computation nearest-neighbour field L1 norm optical flow model occlusion information |
| url | https://doi.org/10.1049/iet-cvi.2019.0321 |
| work_keys_str_mv | AT congxuanzhang stdcflowlargedisplacementflowfieldestimationusingsimilaritytransformationbaseddensecorrespondence AT zhenchen stdcflowlargedisplacementflowfieldestimationusingsimilaritytransformationbaseddensecorrespondence AT fanxiong stdcflowlargedisplacementflowfieldestimationusingsimilaritytransformationbaseddensecorrespondence AT wenliu stdcflowlargedisplacementflowfieldestimationusingsimilaritytransformationbaseddensecorrespondence AT mingli stdcflowlargedisplacementflowfieldestimationusingsimilaritytransformationbaseddensecorrespondence AT liyuege stdcflowlargedisplacementflowfieldestimationusingsimilaritytransformationbaseddensecorrespondence |