Dasen Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Dasen Xu has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Chi Lin
Dalian University of Technology
1 shared publication - Shijun Yan
Dalian University of Technology, Dalian, China
1 shared publication - Tao Jiang
Massey University, Auckland, NZ
1 shared publication - Yi Wang
Dalian University of Technology, Dalian, CN
1 shared publication - Zhongshan Huang
School of Software, Dalian University of Technology, Dalian, Liaoning, China
1 shared publication
Latest Publications
- Open Access
- 22 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Underwater Aquaculture Object Contour Detection Dataset with Benchmarks
Underwater object contour/boundary detection is a field of increasing interest, driven by its potential applications in intelligent aquaculture farming and fishing. However, the limited availability of training data became a bottleneck for this research. We develop an Underwater Aquaculture Object Contour (UAOC) dataset to confront this challenge. UAOC is a comprehensive collection, meticulously annotated pixel-by-pixel, comprising over 7K instances across 1,766 images. These images represent diverse and complex aquatic scenarios, including challenges such as occlusion, cluttered backgrounds, obstructions, multiple objects, and small objects. In addition, we present UAOC-Net, an innovative contour detection network, as a benchmark for the UAOC data set. UAOC-Net leverages multistage feature enhancement and aggregation techniques to achieve precise and clear contours with less noise in the background. The UAOC-Net model and the UAOC dataset are available at https://github.com/yiwangtz/UAOC.
- 22 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Open AccessUnderwater Aquaculture Object Contour Detection Dataset with Benchmarks
Underwater object contour/boundary detection is a field of increasing interest, driven by its potential applications in intelligent aquaculture farming and fishing. However, the limited availability of training data became a bottleneck for this research. We develop an Underwater Aquaculture Object Contour (UAOC) dataset to confront this challenge. UAOC is a comprehensive collection, meticulously annotated pixel-by-pixel, comprising over 7K instances across 1,766 images. These images represent diverse and complex aquatic scenarios, including challenges such as occlusion, cluttered backgrounds, obstructions, multiple objects, and small objects. In addition, we present UAOC-Net, an innovative contour detection network, as a benchmark for the UAOC data set. UAOC-Net leverages multistage feature enhancement and aggregation techniques to achieve precise and clear contours with less noise in the background. The UAOC-Net model and the UAOC dataset are available at https://github.com/yiwangtz/UAOC.