Progressive self-guided loss for salient object detection
We present a simple yet effective progressive self-guided loss function to facilitate deep learning-based salient object detection (SOD) in images. The saliency maps produced by the most relevant works still suffer from incomplete predictions due to the internal complexity of salient objects. Our pr...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/155736 |
| _version_ | 1826122608016359424 |
|---|---|
| author | Yang, Sheng Lin, Weisi Lin, Guosheng Jiang, Qiuping Liu, Zichuan |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Yang, Sheng Lin, Weisi Lin, Guosheng Jiang, Qiuping Liu, Zichuan |
| author_sort | Yang, Sheng |
| collection | NTU |
| description | We present a simple yet effective progressive self-guided loss function to facilitate deep learning-based salient object detection (SOD) in images. The saliency maps produced by the most relevant works still suffer from incomplete predictions due to the internal complexity of salient objects. Our proposed progressive self-guided loss simulates a morphological closing operation on the model predictions for progressively creating auxiliary training supervisions to step-wisely guide the training process. We demonstrate that this new loss function can guide the SOD model to highlight more complete salient objects step-by-step and meanwhile help to uncover the spatial dependencies of the salient object pixels in a region growing manner. Moreover, a new feature aggregation module is proposed to capture multi-scale features and aggregate them adaptively by a branch-wise attention mechanism. Benefiting from this module, our SOD framework takes advantage of adaptively aggregated multi-scale features to locate and detect salient objects effectively. Experimental results on several benchmark datasets show that our loss function not only advances the performance of existing SOD models without architecture modification but also helps our proposed framework to achieve state-of-the-art performance. |
| first_indexed | 2024-10-01T05:51:15Z |
| format | Journal Article |
| id | ntu-10356/155736 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T05:51:15Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | ntu-10356/1557362022-07-21T06:54:30Z Progressive self-guided loss for salient object detection Yang, Sheng Lin, Weisi Lin, Guosheng Jiang, Qiuping Liu, Zichuan School of Computer Science and Engineering School of Electrical and Electronic Engineering Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Salient Object Detection Deep Learning We present a simple yet effective progressive self-guided loss function to facilitate deep learning-based salient object detection (SOD) in images. The saliency maps produced by the most relevant works still suffer from incomplete predictions due to the internal complexity of salient objects. Our proposed progressive self-guided loss simulates a morphological closing operation on the model predictions for progressively creating auxiliary training supervisions to step-wisely guide the training process. We demonstrate that this new loss function can guide the SOD model to highlight more complete salient objects step-by-step and meanwhile help to uncover the spatial dependencies of the salient object pixels in a region growing manner. Moreover, a new feature aggregation module is proposed to capture multi-scale features and aggregate them adaptively by a branch-wise attention mechanism. Benefiting from this module, our SOD framework takes advantage of adaptively aggregated multi-scale features to locate and detect salient objects effectively. Experimental results on several benchmark datasets show that our loss function not only advances the performance of existing SOD models without architecture modification but also helps our proposed framework to achieve state-of-the-art performance. Ministry of Education (MOE) Submitted/Accepted version This work was supported by the Singapore Ministry of Education Tier-2 Fund under Grant MOE2016-T2-2-057(S). 2022-03-21T02:51:26Z 2022-03-21T02:51:26Z 2021 Journal Article Yang, S., Lin, W., Lin, G., Jiang, Q. & Liu, Z. (2021). Progressive self-guided loss for salient object detection. IEEE Transactions On Image Processing, 30, 8426-8438. https://dx.doi.org/10.1109/TIP.2021.3113794 1057-7149 https://hdl.handle.net/10356/155736 10.1109/TIP.2021.3113794 30 8426 8438 en MOE2016-T2-2-057(S) IEEE Transactions on Image Processing © 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/TIP.2021.3113794. application/pdf |
| spellingShingle | Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Salient Object Detection Deep Learning Yang, Sheng Lin, Weisi Lin, Guosheng Jiang, Qiuping Liu, Zichuan Progressive self-guided loss for salient object detection |
| title | Progressive self-guided loss for salient object detection |
| title_full | Progressive self-guided loss for salient object detection |
| title_fullStr | Progressive self-guided loss for salient object detection |
| title_full_unstemmed | Progressive self-guided loss for salient object detection |
| title_short | Progressive self-guided loss for salient object detection |
| title_sort | progressive self guided loss for salient object detection |
| topic | Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision Salient Object Detection Deep Learning |
| url | https://hdl.handle.net/10356/155736 |
| work_keys_str_mv | AT yangsheng progressiveselfguidedlossforsalientobjectdetection AT linweisi progressiveselfguidedlossforsalientobjectdetection AT linguosheng progressiveselfguidedlossforsalientobjectdetection AT jiangqiuping progressiveselfguidedlossforsalientobjectdetection AT liuzichuan progressiveselfguidedlossforsalientobjectdetection |