Deep-learning-based image segmentation integrated with optical microscopy for automatically searching for two-dimensional materials

Abstract

Deep-learning algorithms enable precise image recognition based on high-dimensional hierarchical image features. Here, we report the development and implementation of a deeplearning-based image segmentation algorithm in an autonomous robotic system to search for two-dimensional (2D) materials. We trained the neural network based on Mask-RCNN on annotated optical microscope images of 2D materials (graphene, hBN, MoS2, and WTe2). The inference algorithm is run on a 1024 × 1024 px 2 optical microscope images for 200 ms, enabling the real-time detection of 2D materials. The detection process is robust against changes in the microscopy conditions, such as illumination and color balance, which obviates the parameter-tuning process required for conventional rule-based detection algorithms.Integrating the algorithm with a motorized optical microscope enables the automated searching and cataloging of 2D materials. This development will allow researchers to utilize unlimited amounts of 2D materials simply by exfoliating and running the automated searching process.

RETRACTED: Artificial Intelligence Algorithm Enabled Industrial-Scale Graphene Characterization

“…In recent years, artificial intelligence (AI) tools such as machine-learning paradigms have been proposed to improve the yield of nanomaterials characterization methods [16,17]. For 2D material research, it has been reported that machine learning algorithms integrated with optical microscopy can be used to quantify thickness, impurities, and stacking order in mechanically exfoliated graphene and transition metal chalcogenides [18], and even automatically locate them [19]. Machine learning algorithms have also been used to identify atomic species and defects in transitioning metal chalcogenides by processing high-resolution scanning transmission electron microscopy images [20].…”

Fully Automated Identification of Two-Dimensional Material Samples

“…G. and B. K. contributed equally to this work. Note added.-Recently, we became aware of a similar independent work by Masubuchi et al [45].…”

Machine Learning Analysis of Raman Spectra of MoS2

“…Lin et al first used the support vector machine (SVM) method to learn the contrast information of optical images to determine the layer numbers of graphene and MoS 2 [23]. Later, clustering analysis [24] and the convolutional neural network (CNN) [25][26][27] also joined in this stage play, expanding the identification types and application scenarios of 2D materials. However, because the accuracy of using optical images to determine the layer number based on machine learning is not too high, the previous researchers mainly regarded it as an initial screening to reduce manual work [25].…”