Recent progress in material data mining has been driven by high-capacity models trained on large datasets. However, collecting experimental data (real data) has been extremely costly owing to the amount of human effort and expertise required. Here, we develop a novel transfer learning strategy to address problems of small or insufficient data. This strategy realizes the fusion of real and simulated data and the augmentation of training data in a data mining procedure. For a specific task of grain instance image segmentation, this strategy aims to generate synthetic data by fusing the images obtained from simulating the physical mechanism of grain formation and the “image style” information in real images. The results show that the model trained with the acquired synthetic data and only 35% of the real data can already achieve competitive segmentation performance of a model trained on all of the real data. Because the time required to perform grain simulation and to generate synthetic data are almost negligible as compared to the effort for obtaining real data, our proposed strategy is able to exploit the strong prediction power of deep learning without significantly increasing the experimental burden of training data preparation.

The inner structure of a material is called its microstructure. It stores the genesis of a material and determines all the physical and chemical properties. However, the microstructure is highly complex and numerous image defects such as vague or missing boundaries formed during sample preparation, which makes it difficult to extract the grain boundaries precisely. In this work, we address the task of grain boundary detection in microscopic image processing and develop a graph-cut based method called Fast-FineCut to solve the problem. Our algorithm makes two key contributions: (1) An improved approach that incorporates 3D information between slices as domain knowledge, which can detect the boundaries precisely, even for the vague and missing boundaries. (2) A local processing method based on overlap-tile strategy, which can not only solve the "chain scission" problem at the edge of images, but also economize on the consumption of computing resources. We conduct experiments on a stack of 296 slices of microscopic images of polycrystalline iron (1600 × 2800) and compare the performance against several state-of-the-art boundary detection methods. We conclude that Fast-FineCut can detect boundaries effectively and efficiently.

In material research, it is often highly desirable to observe images of whole microscopic sections with high resolution. So that micrograph stitching is an important technology to produce a panorama or larger image by combining multiple images with overlapping areas, while retaining microscopic resolution. However, due to high complexity and variety of microstructure, most traditional methods could not balance speed and accuracy of stitching strategy. To overcome this problem, we develop a method named very fast sequential micrograph stitching (VFSMS), which employ incremental searching strategy and GPU acceleration to guarantee the accuracy and the speed of stitching results. Experimental results demonstrate that the VFSMS achieve state-of-art performance on three types' microscopic datasets on both accuracy and speed aspects. Besides, it significantly outperforms the most famous and commonly used software, such as ImageJ, Photoshop and Autostitch. The software is available at https://www.mgedata.cn/app_entrance/microscope.

No abstract