Machine Learning-based Automatic Graphene Detection with Color Correction for Optical Microscope Images

Abstract

Graphene serves critical application and research purposes in various fields. However, fabricating highquality and large quantities of graphene is time-consuming and it requires heavy human resource labor costs. In this paper, we propose a Machine Learning-based Automatic Graphene Detection Method with Color Correction (MLA-GDCC), a reliable and autonomous graphene detection from microscopic images. The MLA-GDCC includes a white balance (WB) to correct the color imbalance on the images, a modified U-Net and a support vector machine (SVM) to segment the graphene flakes. Considering the color shifts of the images caused by different cameras, we apply WB correction to correct the imbalance of the color pixels. A modified U-Net model, a convolutional neural network (CNN) architecture for fast and precise image segmentation, is introduced to segment the graphene flakes from the background. In order to improve the pixel-level accuracy, we implement a SVM after the modified U-Net model to separate the monolayer and bilayer graphene flakes. The MLA-GDCC achieves flake-level detection rates of 87.09% for monolayer and 90.41% for bilayer graphene, and the pixel-level accuracy of 99.27% for monolayer and 98.92% for bilayer graphene. MLA-GDCC not only achieves high detection rates of the graphene flakes but also speeds up the latency for the graphene detection process from hours to seconds.

Pixel-wise classification in graphene-detection with tree-based machine learning algorithms

“…This inherent limitation can lead to suppressed F1 score by wrongly identifying the background pixels around the edges of graphene as graphene itself. However, it has been noted that ML models work well for finding graphene flakes even when the F1 score is less than 80% [46]. In figure 5, our DT model with the lowest F1 score of 75.36% among our ML models also finds the graphene flakes well.…”

“…Recent works have used ML techniques to identify the number of layers in a thin film of materials. These researches have employed supervised learning such as support vector machines [39,40], deep neural network (DNN) [41][42][43], U-Net which belongs to CNN [44][45][46], or unsupervised learning such as clustering [41]. The image classifications in the earlier works, however, require too many images for the training dataset that need to be labeled accordingly with layer number, for example, ∼10 3 -10 5 labeled images for DNN [41][42][43], and ∼10 2 [45,46] or 10 3 labeled images for U-Net which are augmented from less than 50 labeled images [44], and more than a dozen GB of GPU memory to process a batch of image data.…”

An open-source robust machine learning platform for real-time detection and classification of 2D material flakes

“…The performance of our model has been evaluated in terms of the IoU, precision, and recall on a pixel and an instance level for various materials and thicknesses. The (pixel-wise determined) recall has been shown to be as high or surpassing those reported for implementations using neural networks [22,44,45]. While the high average recall demonstrates that essentially all flakes are found, the detection algorithm still wrongly detects contamination or shadows, in particular for materials of low optical contrast.…”

Pixel-wise classification in graphene-detection with tree-based machine learning algorithms

“…Recent works have used ML techniques to identify the number of layers in a thin film of materials. These researches have employed supervised learning such as support vector machines (SVM) [34,35], deep neural network (DNN) [36][37][38], U-Net which belongs to CNN [39][40][41], or unsupervised learning such as clustering [36]. The image classifications in the earlier works, however, require too many images for the training dataset that need to be labeled accordingly with layer number, for example, ∼ 10 3 − 10 5 labeled images for DNN [36][37][38], and ∼ 10 2 [40,41] or 10 3 labeled images for U-Net which are augmented from less than 50 labeled images [39], and more than a dozen GB of GPU memory to process a batch of image data.…”