Semantic Segmentation Based on Depth Background Blur
Deep convolutional neural networks (CNNs) are effective in image classification, and are widely used in image segmentation tasks. Several neural netowrks have achieved high accuracy in segementation on existing semantic datasets, for instance PASCAL VOC, CamVid, and Cityscapes. However, there are ne...
| Główni autorzy: | , , |
|---|---|
| Format: | Artykuł |
| Język: | English |
| Wydane: |
MDPI AG
2022-01-01
|
| Seria: | Applied Sciences |
| Hasła przedmiotowe: | |
| Dostęp online: | https://www.mdpi.com/2076-3417/12/3/1051 |
| _version_ | 1827661782952443904 |
|---|---|
| author | Hao Li Changjiang Liu Anup Basu |
| author_facet | Hao Li Changjiang Liu Anup Basu |
| author_sort | Hao Li |
| collection | DOAJ |
| description | Deep convolutional neural networks (CNNs) are effective in image classification, and are widely used in image segmentation tasks. Several neural netowrks have achieved high accuracy in segementation on existing semantic datasets, for instance PASCAL VOC, CamVid, and Cityscapes. However, there are nearly no studies on semantic segmentation from the perspective of a dataset itself. In this paper, we analyzed the characteristics of datasets, and proposed a novel experimental strategy based on bokeh to weaken the impact of futile background information. This crucial bokeh module processed each image in the inference phase by selecting an opportune fuzzy factor <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>σ</mi></semantics></math></inline-formula>, so that the attention of our network can focus on the categories of interest. Some networks based on fully convolutional networks (FCNs) were utilized to verify the effectiveness of our method. Extensive experiments demonstrate that our approach can generally improve the segmentation results on existing datasets, such as PASCAL VOC 2012 and CamVid. |
| first_indexed | 2024-03-10T00:16:01Z |
| format | Article |
| id | doaj.art-7a25b8740a964fc28dc3a4122080b566 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T00:16:01Z |
| publishDate | 2022-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-7a25b8740a964fc28dc3a4122080b5662023-11-23T15:51:05ZengMDPI AGApplied Sciences2076-34172022-01-01123105110.3390/app12031051Semantic Segmentation Based on Depth Background BlurHao Li0Changjiang Liu1Anup Basu2School of Mathematics and Statistics, Sichuan University of Science and Engineering, Zigong 643000, ChinaSchool of Mathematics and Statistics, Sichuan University of Science and Engineering, Zigong 643000, ChinaDepartment of Computing Science, University of Alberta, Edmonton, AB T6G 2H1, CanadaDeep convolutional neural networks (CNNs) are effective in image classification, and are widely used in image segmentation tasks. Several neural netowrks have achieved high accuracy in segementation on existing semantic datasets, for instance PASCAL VOC, CamVid, and Cityscapes. However, there are nearly no studies on semantic segmentation from the perspective of a dataset itself. In this paper, we analyzed the characteristics of datasets, and proposed a novel experimental strategy based on bokeh to weaken the impact of futile background information. This crucial bokeh module processed each image in the inference phase by selecting an opportune fuzzy factor <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mi>σ</mi></semantics></math></inline-formula>, so that the attention of our network can focus on the categories of interest. Some networks based on fully convolutional networks (FCNs) were utilized to verify the effectiveness of our method. Extensive experiments demonstrate that our approach can generally improve the segmentation results on existing datasets, such as PASCAL VOC 2012 and CamVid.https://www.mdpi.com/2076-3417/12/3/1051bokehfully convolutional networkssemantic segmentation |
| spellingShingle | Hao Li Changjiang Liu Anup Basu Semantic Segmentation Based on Depth Background Blur Applied Sciences bokeh fully convolutional networks semantic segmentation |
| title | Semantic Segmentation Based on Depth Background Blur |
| title_full | Semantic Segmentation Based on Depth Background Blur |
| title_fullStr | Semantic Segmentation Based on Depth Background Blur |
| title_full_unstemmed | Semantic Segmentation Based on Depth Background Blur |
| title_short | Semantic Segmentation Based on Depth Background Blur |
| title_sort | semantic segmentation based on depth background blur |
| topic | bokeh fully convolutional networks semantic segmentation |
| url | https://www.mdpi.com/2076-3417/12/3/1051 |
| work_keys_str_mv | AT haoli semanticsegmentationbasedondepthbackgroundblur AT changjiangliu semanticsegmentationbasedondepthbackgroundblur AT anupbasu semanticsegmentationbasedondepthbackgroundblur |