中国物理B ›› 2024, Vol. 33 ›› Issue (3): 30703-030703.doi: 10.1088/1674-1056/ad23d8

• • 上一篇    下一篇

Image segmentation of exfoliated two-dimensional materials by generative adversarial network-based data augmentation

Xiaoyu Cheng(程晓昱)1, Chenxue Xie(解晨雪)1, Yulun Liu(刘宇伦)1, Ruixue Bai(白瑞雪)1, Nanhai Xiao(肖南海)1, Yanbo Ren(任琰博)1, Xilin Zhang(张喜林)1, Hui Ma(马惠)2,†, and Chongyun Jiang(蒋崇云)1,‡   

  1. 1 College of Electronic Information and Optical Engineering, Nankai University, Tianjin 300350, China;
    2 School of Physical Science and Technology, Tiangong University, Tianjin 300387, China
  • 收稿日期:2024-01-02 修回日期:2024-01-26 接受日期:2024-01-30 出版日期:2024-02-22 发布日期:2024-03-06
  • 通讯作者: Hui Ma, Chongyun Jiang E-mail:mahuimoving@163.com;jiang.chongyun@nankai.edu.cn
  • 基金资助:
    Project supported by the National Key Research and Development Program of China (Grant No. 2022YFB2803900), the National Natural Science Foundation of China (Grant Nos. 61974075 and 61704121), the Natural Science Foundation of Tianjin Municipality (Grant Nos. 22JCZDJC00460 and 19JCQNJC00700), Tianjin Municipal Education Commission (Grant No. 2019KJ028), and Fundamental Research Funds for the Central Universities (Grant No. 22JCZDJC00460).

Image segmentation of exfoliated two-dimensional materials by generative adversarial network-based data augmentation

Xiaoyu Cheng(程晓昱)1, Chenxue Xie(解晨雪)1, Yulun Liu(刘宇伦)1, Ruixue Bai(白瑞雪)1, Nanhai Xiao(肖南海)1, Yanbo Ren(任琰博)1, Xilin Zhang(张喜林)1, Hui Ma(马惠)2,†, and Chongyun Jiang(蒋崇云)1,‡   

  1. 1 College of Electronic Information and Optical Engineering, Nankai University, Tianjin 300350, China;
    2 School of Physical Science and Technology, Tiangong University, Tianjin 300387, China
  • Received:2024-01-02 Revised:2024-01-26 Accepted:2024-01-30 Online:2024-02-22 Published:2024-03-06
  • Contact: Hui Ma, Chongyun Jiang E-mail:mahuimoving@163.com;jiang.chongyun@nankai.edu.cn
  • Supported by:
    Project supported by the National Key Research and Development Program of China (Grant No. 2022YFB2803900), the National Natural Science Foundation of China (Grant Nos. 61974075 and 61704121), the Natural Science Foundation of Tianjin Municipality (Grant Nos. 22JCZDJC00460 and 19JCQNJC00700), Tianjin Municipal Education Commission (Grant No. 2019KJ028), and Fundamental Research Funds for the Central Universities (Grant No. 22JCZDJC00460).

摘要: Mechanically cleaved two-dimensional materials are random in size and thickness. Recognizing atomically thin flakes by human experts is inefficient and unsuitable for scalable production. Deep learning algorithms have been adopted as an alternative, nevertheless a major challenge is a lack of sufficient actual training images. Here we report the generation of synthetic two-dimensional materials images using StyleGAN3 to complement the dataset. DeepLabv3Plus network is trained with the synthetic images which reduces overfitting and improves recognition accuracy to over 90%. A semi-supervisory technique for labeling images is introduced to reduce manual efforts. The sharper edges recognized by this method facilitate material stacking with precise edge alignment, which benefits exploring novel properties of layered-material devices that crucially depend on the interlayer twist-angle. This feasible and efficient method allows for the rapid and high-quality manufacturing of atomically thin materials and devices.

关键词: two-dimensional materials, deep learning, data augmentation, generating adversarial networks

Abstract: Mechanically cleaved two-dimensional materials are random in size and thickness. Recognizing atomically thin flakes by human experts is inefficient and unsuitable for scalable production. Deep learning algorithms have been adopted as an alternative, nevertheless a major challenge is a lack of sufficient actual training images. Here we report the generation of synthetic two-dimensional materials images using StyleGAN3 to complement the dataset. DeepLabv3Plus network is trained with the synthetic images which reduces overfitting and improves recognition accuracy to over 90%. A semi-supervisory technique for labeling images is introduced to reduce manual efforts. The sharper edges recognized by this method facilitate material stacking with precise edge alignment, which benefits exploring novel properties of layered-material devices that crucially depend on the interlayer twist-angle. This feasible and efficient method allows for the rapid and high-quality manufacturing of atomically thin materials and devices.

Key words: two-dimensional materials, deep learning, data augmentation, generating adversarial networks

中图分类号:  (Image processing)

  • 07.05.Pj
68.65.-k (Low-dimensional, mesoscopic, nanoscale and other related systems: structure and nonelectronic properties) 84.35.+i (Neural networks) 87.64.M- (Optical microscopy)