Automatic detection of multilayer hexagonal boron nitride in optical images using deep learning-based computer vision

Abstract

Computer vision algorithms can quickly analyze numerous images and identify useful information with high accuracy. Recently, computer vision has been used to identify 2D materials in microscope images. 2D materials have important fundamental properties allowing for their use in many potential applications, including many in quantum information science and engineering. One such material is hexagonal boron nitride (hBN), an isomorph of graphene with a very indistinguishable layered structure. In order to use these materials for research and product development, the most effective method is mechanical exfoliation where single-layer 2D crystallites must be prepared through an exfoliation procedure and then identified using reflected light optical microscopy. Performing these searches manually is a time-consuming and tedious task. Deploying deep learning-based computer vision algorithms for 2D material search can automate the flake detection task with minimal need for human intervention. In this work, we have implemented a new deep learning pipeline to classify crystallites of hBN based on coarse thickness classifications in reflected-light optical micrographs. We have used DetectoRS as the object detector and trained it on 177 images containing hexagonal boron nitride (hBN) flakes of varying thickness. The trained model achieved a high detection accuracy for the rare category of thin flakes ($$<50$$ < 50 atomic layers thick). Further analysis shows that our proposed pipeline could be generalized to various microscope settings and is robust against changes in color or substrate background.

Topological data analysis enhanced prediction of hydrogen storage in metal–organic frameworks (MOFs)

“…Examples include the ability of computer vision algorithms to swiftly and accurately analyse large numbers of images and extract relevant data. 35 Since TDA involves visual data, using deep learning-based computer vision algorithms to examine the outcomes seems like a natural fit. This line of thinking led us to consider implementing a cutting-edge CNN-based architecture called residual network (ResNet), which was initially developed by a team of Microsoft researchers and has since shown significant gains in a number of different settings.…”

TLN-LSTM: an automatic modulation recognition classifier based on a two-layer nested structure of LSTM network for extremely long signal sequences

“…Traditional modulation identification methods are highly dependent on prior knowledge, have the disadvantages of high computational complexity and low identification accuracy, and are difficult to apply in practical problems (Xu et al , 2010). In recent years, deep learning has shined in the fields of natural language processing (Otter et al , 2020), computer vision (Ramezani et al , 2023), speech recognition (Weng et al , 2023) and information retrieval (Liang et al , 2023), and has gradually attracted attention from all walks of life. To overcome the inherent shortcomings of traditional modulation recognition algorithms, researchers in the field of wireless communications have also begun to apply deep learning ideas to the field of signal modulation recognition, which can often achieve better recognition performance.…”

A new detection algorithm for alien intrusion on highway

“…In the past decade, with the rapid development of deep learning, object detection algorithms based on deep learning have been widely recognized in various fields 2 . The complex environment on the highway is easily affected by light and shadow changes, as well as the large number of vehicles and the subjective influence of human eyes, which makes it difficult for traditional cameras to directly capture foreign objects.…”