在金属图像中进行密集预测预训的补丁样本对比学习
Mingchun Li1,2, Yang Liu3, Dali Chen4
1School of Intelligent Science and Information Engineering, Shenyang University, Shenyang, 110044, China. limingchun_cn@qq.com.
Scientific reports
|December 16, 2025
概括
这项研究引入了一种新的补丁采样对比学习 (PSCL) 方法,用于金属图像中的微结构细分. PSCL有效地捕获全球和本地特征,显著提高了对细分的准确性,使用最小的注释数据.
科学领域:
- 材料科学 材料科学 材料科学
- 计算机视觉 计算机视觉
- 机器学习 机器学习
背景情况:
- 微观结构特征对于合金的机械性能至关重要.
- 深度学习在图像分析方面表现出色,但受到高标注成本的阻碍.
- 现有的自我监督学习方法需要适应微观结构特定的任务.
研究的目的:
- 开发一个特定于微观结构的预培训框架,以解决注释成本.
- 增强基于深度学习的金属图像中的微结构识别.
- 为了提高微结构细分的效率和准确性.
主要方法:
- 提出了一种新的补丁样本对比学习 (PSCL) 方法.
- 实现图像级和补丁级对比学习,用于全球和本地特征捕获.
- 引入了多尺度策略和基于特征相似性的采样方法,以提高适应性和可区分性.
主要成果:
- 在微调后只用一个注释图像实现了0.6296的子得分.
- 优于现有的自我监督学习方法,具有相同的模型结构.
- 证明了PSCL在微结构细分中的有效性.
结论:
- PSCL是一种高效的自主监督学习方法,用于金属图像细分.
- 该方法显著减少了对广泛数据注释的需求.
- 在制造过程中,PSCL为微结构识别提供了一个有前途的解决方案.
相关概念视频
01:20Preparation of Samples for Electron Microscopy
6.7K
To be visualized by an electron microscope, either transmission or scanning, biological samples need to be fixed (stabilized) so the electron beam does not destroy them and dried thoroughly (desiccated/dehydrated) so the vacuum does not affect them. Fixation needs to be done as quickly as possible because the sample properties will start changing as soon as it is removed from its natural environment. For example, in a tissue sample, the oxygen levels begin decreasing, causing an altered...
6.7K
01:05Difference from Background: Limit of Detection
8.0K
The limit of detection (LOD) is the smallest amount of analyte that can be distinguished from the background noise. The LOD value corresponds to the concentration at which the analyte signal is three times larger than the standard deviation of the blank signal. Below this value, the analyte signal cannot be differentiated from the background noise. It is calculated by dividing the calibration slope by 3 times the standard deviation of the blank signals.
The LOD indicates the presence or absence...
The LOD indicates the presence or absence...
8.0K