RGB-D salient object detection based on BC2 FNet network
In the face of complex scene images, the introduction of depth information can greatly improve the performance of salient object detection. However, up-sampling and down-sampling operations in neural networks maybe blur the boundaries of objects in the saliency map, thereby reducing the performance...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
EDP Sciences
2024-12-01
|
| Series: | Xibei Gongye Daxue Xuebao |
| Subjects: | |
| Online Access: | https://www.jnwpu.org/articles/jnwpu/full_html/2024/06/jnwpu2024426p1135/jnwpu2024426p1135.html |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1825206546455330816 |
|---|---|
| author | WANG Feng CHENG Yongmei |
| author_facet | WANG Feng CHENG Yongmei |
| author_sort | WANG Feng |
| collection | DOAJ |
| description | In the face of complex scene images, the introduction of depth information can greatly improve the performance of salient object detection. However, up-sampling and down-sampling operations in neural networks maybe blur the boundaries of objects in the saliency map, thereby reducing the performance of salient object detection. Aiming at this problem, a boundary-driven cross-modal and cross-layer fusion network (BC2FNet) for RGB-D salient object detection is proposed in this paper, which preserves the boundary of the object by adding the guidance of boundary information to the cross-modal and cross-layer fusion, respectively. Firstly, a boundary generation module is designed to extract two kinds of boundary information from low-level features of RGB and depth modalities, respectively. Secondly, a boundary-driven feature selection module is designed, which is dedicated to simultaneously focusing on important feature information and preserving boundary details in the process of RGB and depth modality fusion. Finally, a boundary-driven cross-layer fusion module is proposed which simultaneously adds two kinds of boundary information in the process of up-sampling fusion on adjacent layers. By embedding this module into the top-down information fusion flow, the predicted saliency map can contain accurate objects and sharp boundaries. Simulation results on five standard RGB-D data sets show that the proposed model can achieve better performance. |
| format | Article |
| id | doaj-art-1d4a45d0f5f34abca12dc0331b52edf8 |
| institution | Kabale University |
| issn | 1000-2758 2609-7125 |
| language | zho |
| publishDate | 2024-12-01 |
| publisher | EDP Sciences |
| record_format | Article |
| series | Xibei Gongye Daxue Xuebao |
| spelling | doaj-art-1d4a45d0f5f34abca12dc0331b52edf82025-02-07T08:23:13ZzhoEDP SciencesXibei Gongye Daxue Xuebao1000-27582609-71252024-12-014261135114310.1051/jnwpu/20244261135jnwpu2024426p1135RGB-D salient object detection based on BC2 FNet networkWANG Feng0CHENG Yongmei1School of Physics and Electrical Engineering, Weinan Normal UniversitySchool of Automation, Northwestern Polytechnical UniversityIn the face of complex scene images, the introduction of depth information can greatly improve the performance of salient object detection. However, up-sampling and down-sampling operations in neural networks maybe blur the boundaries of objects in the saliency map, thereby reducing the performance of salient object detection. Aiming at this problem, a boundary-driven cross-modal and cross-layer fusion network (BC2FNet) for RGB-D salient object detection is proposed in this paper, which preserves the boundary of the object by adding the guidance of boundary information to the cross-modal and cross-layer fusion, respectively. Firstly, a boundary generation module is designed to extract two kinds of boundary information from low-level features of RGB and depth modalities, respectively. Secondly, a boundary-driven feature selection module is designed, which is dedicated to simultaneously focusing on important feature information and preserving boundary details in the process of RGB and depth modality fusion. Finally, a boundary-driven cross-layer fusion module is proposed which simultaneously adds two kinds of boundary information in the process of up-sampling fusion on adjacent layers. By embedding this module into the top-down information fusion flow, the predicted saliency map can contain accurate objects and sharp boundaries. Simulation results on five standard RGB-D data sets show that the proposed model can achieve better performance.https://www.jnwpu.org/articles/jnwpu/full_html/2024/06/jnwpu2024426p1135/jnwpu2024426p1135.htmlsalient object detectionboundary-drivencross-modal fusioncross-layer fusion |
| spellingShingle | WANG Feng CHENG Yongmei RGB-D salient object detection based on BC2 FNet network Xibei Gongye Daxue Xuebao salient object detection boundary-driven cross-modal fusion cross-layer fusion |
| title | RGB-D salient object detection based on BC2 FNet network |
| title_full | RGB-D salient object detection based on BC2 FNet network |
| title_fullStr | RGB-D salient object detection based on BC2 FNet network |
| title_full_unstemmed | RGB-D salient object detection based on BC2 FNet network |
| title_short | RGB-D salient object detection based on BC2 FNet network |
| title_sort | rgb d salient object detection based on bc2 fnet network |
| topic | salient object detection boundary-driven cross-modal fusion cross-layer fusion |
| url | https://www.jnwpu.org/articles/jnwpu/full_html/2024/06/jnwpu2024426p1135/jnwpu2024426p1135.html |
| work_keys_str_mv | AT wangfeng rgbdsalientobjectdetectionbasedonbc2fnetnetwork AT chengyongmei rgbdsalientobjectdetectionbasedonbc2fnetnetwork |