A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons

Three-dimensional (3D) data analysis, especially 3D reconstruction, is critical within integrated circuits for quality control and inspection. The traditional 3D reconstruction, like multi-energy deconvolution scanning electron microscopy (MED-SEM), can achieved resolution of 5 nm in the horizontal...

Full description

Saved in:
Bibliographic Details
Main Authors: Caizhi Zheng, Ronghan Hong, Hao-Jie Hu, Qing Huo Liu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11045923/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849433981319118848
author Caizhi Zheng
Ronghan Hong
Hao-Jie Hu
Qing Huo Liu
author_facet Caizhi Zheng
Ronghan Hong
Hao-Jie Hu
Qing Huo Liu
author_sort Caizhi Zheng
collection DOAJ
description Three-dimensional (3D) data analysis, especially 3D reconstruction, is critical within integrated circuits for quality control and inspection. The traditional 3D reconstruction, like multi-energy deconvolution scanning electron microscopy (MED-SEM), can achieved resolution of 5 nm in the horizontal direction and 10 nm in the vertical direction from backscattered electron images of objects at different depth layers. However, the traditional methods are difficult to obtain high-precision reconstruction results in the longitudinal direction, and the reconstruction results of most methods contain a large number of artifacts. Thus, in this work, a hybrid CNN-Transformer network, called BSE-VoxNets, is proposed for 3D high precision reconstruction. The proposed BSE-VoxNets consists three parts: transformer feature extractor, convolutional upsampling block and feature fusion block. 10000 BSE samples were used for training BSE-VoxNets, with 200 samples for testing, and the optimal cross-union ratio (IoU) of the test dataset reached 0.953. Two numerical cases are used to respectively verify the inversion performance of the proposed method for different depths and different feature scales. The reconstruction results show that the proposed BSE-VoxNets can reach a resolution of 2 nm in both the horizontal and vertical dimensions.
format Article
id doaj-art-4e8c995ef2b64253a4cd9109ee640e5b
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-4e8c995ef2b64253a4cd9109ee640e5b2025-08-20T03:26:49ZengIEEEIEEE Access2169-35362025-01-011310855110856010.1109/ACCESS.2025.358197411045923A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered ElectronsCaizhi Zheng0https://orcid.org/0009-0003-3424-2171Ronghan Hong1Hao-Jie Hu2https://orcid.org/0009-0004-3189-1552Qing Huo Liu3https://orcid.org/0000-0001-5286-4423Institute of Electromagnetics and Acoustics, Fujian Provincial Key Laboratory of Electromagnetic Wave Science and Detection Technology, Xiamen University, Xiamen, ChinaTangshan Technology Company, Ningbo, ChinaInstitute of Electromagnetics and Acoustics, Fujian Provincial Key Laboratory of Electromagnetic Wave Science and Detection Technology, Xiamen University, Xiamen, ChinaInstitute of Electromagnetics and Acoustics, Fujian Provincial Key Laboratory of Electromagnetic Wave Science and Detection Technology, Xiamen University, Xiamen, ChinaThree-dimensional (3D) data analysis, especially 3D reconstruction, is critical within integrated circuits for quality control and inspection. The traditional 3D reconstruction, like multi-energy deconvolution scanning electron microscopy (MED-SEM), can achieved resolution of 5 nm in the horizontal direction and 10 nm in the vertical direction from backscattered electron images of objects at different depth layers. However, the traditional methods are difficult to obtain high-precision reconstruction results in the longitudinal direction, and the reconstruction results of most methods contain a large number of artifacts. Thus, in this work, a hybrid CNN-Transformer network, called BSE-VoxNets, is proposed for 3D high precision reconstruction. The proposed BSE-VoxNets consists three parts: transformer feature extractor, convolutional upsampling block and feature fusion block. 10000 BSE samples were used for training BSE-VoxNets, with 200 samples for testing, and the optimal cross-union ratio (IoU) of the test dataset reached 0.953. Two numerical cases are used to respectively verify the inversion performance of the proposed method for different depths and different feature scales. The reconstruction results show that the proposed BSE-VoxNets can reach a resolution of 2 nm in both the horizontal and vertical dimensions.https://ieeexplore.ieee.org/document/11045923/3D reconstructiondeep learningmachine learningscanning electron microscopy
spellingShingle Caizhi Zheng
Ronghan Hong
Hao-Jie Hu
Qing Huo Liu
A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons
IEEE Access
3D reconstruction
deep learning
machine learning
scanning electron microscopy
title A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons
title_full A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons
title_fullStr A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons
title_full_unstemmed A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons
title_short A Voxelized Transformer-Based Neural Network for 3D Reconstruction From Multi-Energy SEM Backscattered Electrons
title_sort voxelized transformer based neural network for 3d reconstruction from multi energy sem backscattered electrons
topic 3D reconstruction
deep learning
machine learning
scanning electron microscopy
url https://ieeexplore.ieee.org/document/11045923/
work_keys_str_mv AT caizhizheng avoxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT ronghanhong avoxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT haojiehu avoxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT qinghuoliu avoxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT caizhizheng voxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT ronghanhong voxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT haojiehu voxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons
AT qinghuoliu voxelizedtransformerbasedneuralnetworkfor3dreconstructionfrommultienergysembackscatteredelectrons