Identification and classification of exfoliated graphene flakes from microscopy images using a hierarchical deep convolutional neural network
Identification of the mechanically exfoliated graphene flakes and classification of the thickness is important in the nanomanufacturing of next-generation materials and devices that overcome the bottleneck of Moore's Law. Currently, identification and classification of exfoliated graphene flakes are conducted by human via inspecting the optical microscope images. The existing state-of-the-art automatic identification by machine learning is not able to accommodate images with different backgrounds while different backgrounds are unavoidable in experiments. This paper presents a deep learning method to automatically identify and classify the thickness of exfoliated graphene flakes on Si/SiO2 substrates from optical microscope images with various settings and background colors. The presented method uses a hierarchical deep convolutional neural network that is capable of learning new images while preserving the knowledge from previous images. The deep learning model was trained and used to classify exfoliated graphene flakes into monolayer (1L), bi-layer (2L), tri-layer (3L), four-to-six-layer (4-6L), seven-to-ten-layer (7-10L), and bulk categories. Compared with existing machine learning methods, the presented method possesses high accuracy and efficiency as well as robustness to the backgrounds and resolutions of images. The results indicated that our deep learning model has accuracy as high as 99% in identifying and classifying exfoliated graphene flakes. This research will shed light on scaled-up manufacturing and characterization of graphene for advanced materials and devices.