期刊
JOURNAL OF COMPUTATIONAL SCIENCE
卷 62, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.jocs.2022.101709
关键词
Material microscopic image segmentation; Convolution neural network; Loss function
资金
- National Key Research and Devel-opment Program of China [2020YFB0704501]
- National Natural Science Foundation of China [62106019, 2021M700383]
- Scientific and Technological Innovation Foundation of Shunde Graduate School of USTB [BK20AF001, BK21BF002]
- Fun-damental Research Funds for the Central Universities of China [00007467]
- Postdoctor Research Foundation of Shunde Grad-uate School of University of Science and Technology Beijing [2021BH005]
- USTB MatCom of Beijing Advanced Innovation Center for Materials Genome Engineer-ing
This paper proposes a novel Weighted Propagation Convolution Neural Network based on U-Net (WPU-Net) for boundary detection in poly-crystalline microscopic images. By introducing spatial consistency and adaptive boundary weight, the method achieves promising performance in image segmentation.
In material science, image segmentation is of great significance for quantitative analysis of microstructures. Here, we propose a novel Weighted Propagation Convolution Neural Network based on U-Net (WPU-Net) to detect boundary in poly-crystalline microscopic images. We introduce spatial consistency into network to eliminate the defects in raw microscopic image. And we customize adaptive boundary weight for each pixel in each grain, so that it leads the network to preserve grain's geometric and topological characteristics. Moreover, we provide our dataset with the goal of advancing the development of image processing in materials science. Experiments demonstrate that the proposed method achieves promising performance in both of objective and subjective assessment. In boundary detection task, it reduces the error rate by 7%, which outperforms state-of-the-art methods by a large margin.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据