A parallel down-up fusion network for salient object detection in optical remote sensing images

NEUROCOMPUTING

卷:415

Article

Li, Chongyi¹  Cong, Runmin^2,3  Guo, Chunle⁴  Li, Hua^5,6  Zhang, Chunjie^2,3  Zheng, Feng⁷  Zhao, Yao^2,3 

[1] Tianjin Univ, Sch Elect & Informat Engn, Tianjin 300072, Peoples R China

[2] Beijing Jiaotong Univ, Inst Informat Sci, Beijing 100044, Peoples R China

[3] Beijing Key Lab Adv Informat Sci & Network Techno, Beijing 100044, Peoples R China

[4] Nankai Univ, Coll Comp Sci, Tianjin 300350, Peoples R China

[5] Huazhong Univ Sci & Technol, Sch Software Engn, Wuhan 430074, Peoples R China

[6] City Univ Hong Kong, Dept Comp Sci, Kowloon, Hong Kong 999077, Peoples R China

[7] Southern Univ Sci & Technol, Dept Comp Sci & Engn, Shenzhen 518055, Peoples R China

关键词: Optical remote sensing images;    Salient object detection;    Deep learning;   

DOI  :  10.1016/j.neucom.2020.05.108

来源: Elsevier

【 摘 要 】

The diverse spatial resolutions, various object types, scales and orientations, and cluttered backgrounds in optical remote sensing images (RSIs) challenge the current salient object detection (SOD) approaches. It is commonly unsatisfactory to directly employ the SOD approaches designed for nature scene images (NSIs) to RSIs. In this paper, we propose a novel Parallel Down-up Fusion network (PDF-Net) for SOD in optical RSIs, which takes full advantage of the in-path low- and high-level features and cross-path multi-resolution features to distinguish diversely scaled salient objects and suppress the cluttered backgrounds. To be specific, keeping a key observation that the salient objects still are salient no matter the resolutions of images are in mind, the PDF-Net takes successive down-sampling to form five parallel paths and perceive scaled salient objects that are commonly existed in optical RSIs. Meanwhile, we adopt the dense connections to take advantage of both low- and high-level information in the same path and build up the relations of cross paths, which explicitly yield strong feature representations. At last, we fuse the multiple-resolution features in parallel paths to combine the benefits of the features with different resolutions, i.e., the high-resolution feature consisting of complete structure and clear details while the low-resolution features highlighting the scaled salient objects. Extensive experiments on the ORSSD dataset demonstrate that the proposed network is superior to the state-of-the-art approaches both qualitatively and quantitatively. (C) 2020 Elsevier B.V. All rights reserved.

【 授权许可】

Free